Spatiotemporal Analysis And Causal Study Of Building Collapse Accidents

In recent years,China has seen a rapid increase in construction enterprises,active projects and an increase in the number of people working in the construction industry.At the same time,construction safety accidents are frequent and easy to occur.According to official statistics,construction collapses are characterised by a "high number" and "high mortality rate",and the safety situation cannot be underestimated.Once a building collapse occurs,it will have irreversible consequences for society and individuals.Therefore,it is of paramount importance for the construction industry to study the characteristics of construction collapse accidents and the factors contributing to them,so as to effectively prevent and control the occurrence of accidents and reduce their losses.In order to improve the safety management of government and enterprises,and to effectively prevent and control construction collapse accidents,this paper aims to explore the spatial and temporal distribution characteristics of construction collapse accidents,and to carry out research on the identification and management of the factors contributing to construction collapse accidents.Based on the theoretical foundation,this paper uses GIS as the basic tool to analyse the characteristics of the spatial and temporal distribution of building collapse accidents based on 594 accident investigation reports collected and collated from official websites from 2000 to 2020.Using descriptive statistics,it was found that the overall trend of building collapse accidents was on the rise,reaching a peak in 2017,followed by a decline in the number of accidents;the peak and trough periods were August and February respectively;the daily periods of 8:00-11:00 and 14:00-17:00 were the most frequent periods of accidents,with 11:00 The two peak periods are 11:00 and 16:00;spatially,using a variety of spatial analysis methods such as imbalance index,kernel density analysis and standard deviation ellipse,it can be seen that building collapse accidents are aggregated and unbalanced in distribution,in line with the characteristics of Hu Huanyong distribution;Guangdong Province and Jiangsu Province are the high accident areas,with an obvious trend of shifting to the northeast,and the agglomeration effect from west to east is gradually increasing,and the area of the occurrence area is shrinking There is a tendency for the occurrence area to shrink.In terms of the identification and modelling of accident causal factors,the results of the spatio-temporal analysis were used as support,and the data from the hotspot areas were used as the basis,supplemented by a small amount of data from secondary areas,to identify the causal factors of construction collapse accidents based on text mining technology,and to improve the causal factor system by combining literature research and expert interviews.The system is based on a combination of literature research and expert interviews,and is based on four factors:personnel,physical,environmental and management factors.20 contributing factors to building collapse accidents were identified,including illegal work,poor safety awareness,geological conditions,quality of construction materials and site safety inspection.Based on this,an expert interview team was formed to construct the adjacency matrix of the factors contributing to the building collapse accidents,solve the reachable matrix using python,and draw the explanatory structure model to lay the foundation for the construction of the Bayesian model later.A questionnaire was distributed to obtain the data,and the data was standardised using the risk matrix.The standardised data was then used to learn the Bayesian network structure with the aid of GeNIe 2.1 Academic tool to form a complete Bayesian network model of the factors contributing to building collapse.Next,Bayesian network parameters were learned.From the sensitivity analysis,we can see that 10 risk factors,including administrative supervision,physical and psychological conditions,climatic conditions,inadequate safety protection measures,responsibility of the main body of safety management,geological conditions,safety awareness,personal safety protection,irregularities,and safety education and training,are the more sensitive factors among the contributing factors of construction collapse accidents;from the most general causal chain,we can see that administrative supervision is the only causal source of construction collapse accidents.Eight causal chains of construction collapse accidents were identified.The accident in the second phase of the Waterfront Qinghua high-rise project in Wuzhong District,Suzhou,was then taken as an example,and the risk level of each factor was comprehensively judged and input into the network model of the building collapse accident,resulting in the conclusion that the building collapse accident had a high probability of occurrence,which was basically consistent with the actual situation and verified the feasibility of the Bayesian network model.Focusing on the accident itself,this paper constructs a relatively comprehensive framework for accident research based on construction collapse accident cases and official accident investigation reports as the data source,analyses the spatial and temporal analysis characteristics of accidents and the relationship between factors,and proposes targeted management countermeasures based on the research results,providing a basis for relevant departments to prevent and control construction collapse accidents and helping to improve safety management.