Research On Influence Of Construction Safety Management Strategy On Safety Performance Based On Bayesian Network

The construction industry is one of the high-risk industries with frequent accidents.The occurrence of safety accidents will not only seriously damage people's lives and property safety,but also affect the stable development of society.As an important influencing factor of safety accidents,safety management measures have been unremittingly explored by scholars and achieved fruitful results.However,little attention has been paid to the impact of the combination of safety management measures on safety performance and the interaction between safety management measures.Therefore,this paper explores the optimal management strategy for improving safety performance by constructing a Bayesian network model between commonly used safety management measures,with a view to providing a theoretical basis for the study of improving safety performance of construction projects.The main work of this paper are divided into the following four parts.First,based on the literature review,combined with the current status of safety management measures for construction projects,after several expert interviews and discussions,a measurement scale for safety management measures and safety performance was constructed,which was verified through reliability and validity analysis.The developed scale has good reliability and validity.Second,use the software SPSS Modeler to perform data mining on the data collected by the scale,and construct a tree extended Bayesian network model(TAN),a Markov blanket network(Markov),and a feature selection Markov blanket network(Markov-FS),And through evaluation and analysis,the tree expansion Bayesian network model has the highest accuracy,so the tree expansion Bayesian network model is selected for empirical research.Then,based on the tree-expanded Bayesian network model,a theoretical hypothesis is proposed,a structural equation model of the impact of safety management measures on the safety performance relationship is constructed,and empirical research is performed using Smart PLS.Analysis to verify proposed hypotheses.The research results show that safety incentives,safety inspections,safety resource supply,worker participation,safety training,accident investigation and analysis,and management commitment significantly affect the safety performance of construction projects.Safety planning,communication and feedback,and safety penalties have no direct impact on safety performance.Significantly,safety performance can be indirectly affected through other safety management measures.Finally,the Bayesian network model was modified based on the test results of Smart PLS,and research and analysis was performed using Netica software.The results show that when the10 safety management measures are implemented well,the probability of the safety performance level is as high as 98.8%;The probability of good performance exceeds 95%.Safety incentives,worker participation,management commitment,accident investigation and analysis,and safety training or safety resource provision,safety incentives,safety training,accident investigation and analysis,and management commitment can be taken.Management measures as the optimal safety management strategy.