Font Size: a A A

Study On Causes Analysis And Prediction Model Of Construction Safety Accidents

Posted on:2022-07-31Degree:MasterType:Thesis
Country:ChinaCandidate:J R ZhangFull Text:PDF
GTID:2492306575974159Subject:Architecture and Civil Engineering
Abstract/Summary:PDF Full Text Request
Since the founding of the People’s Republic of China,China’s construction industry has developed rapidly,the scale of output continues to expand,and has broken through the historical highs one after another.The rapid development of the construction industry is also faced with huge risks and challenges.Due to the large scale of construction projects,high labor intensity,long construction cycle,complex and changeable environment,backward construction management methods,etc.,China’s construction safety production situation is still grim.In order to improve the situation of construction safety accidents,Based on 1325 safety investigation reports of construction safety accidents,this paper uses Logit model and machine learning methods to conduct causation analysis and build a prediction model,which provides reference for the improvement of construction safety management and accident prevention.Firstly,a descriptive statistical analysis of 1325 construction safety accidents was conducted from three aspects of accident levels,accident types,accident time characteristics,revealing the basic rules of the accidents.Then,based on system thinking and accident cause theory,According to the hierarchy of "system-subsystem-factor-sub-factor",the construction safety accident cause system was decomposed into 6 subsystems which include Organizational Management,Technical Management,Working Behavior,Object Control,Contract Management,Environmental Management and 40 subfactors.Secondly,with accident causes as independent variables and accident consequence as dependent variables,ordered Logit and multinomial Logit models were built to explore the nonlinear relationship between causes,accident level and accident type,and the key causes were obtained.Then,based on the random forest algorithm,the construction safety accident prediction model was built.After debugging,the prediction accuracy of accident grade and accident type reached0.91 and 0.78.Finally,the "3·7" collapse accident of Xinjia Hotel in Quanzhou City,Fujian Province was selected as a case to verify the applicability of the model.
Keywords/Search Tags:Construction safety, System thinking, Accident causes, Logit model, Machine learning
PDF Full Text Request
Related items