Research On Unauthorized Intrusion Monitoring And Safety Accident Prediction In Large-Scale Project Construction Site

Large-scale projects,such as water conservancy and hydropower projects,often have large scale,long construction period and great pressure on construction safety management.In the safety management of construction site,accident prevention is the key and the important embodiment of prior control.Unauthorized intrusion is related to multiple accidents such as falling from height,objects striking and electric shock.The monitoring of unauthorized intrusion usually adopts the technologies of human supervision,vidiomoitoring,computer vision,and other ways,which often have the problems that real-time control,accurate inspection and excessive alarm need to be improved.The prediction of unauthorized intrusion related accidents is usually difficult,such as the subjective deviation of expert judgment and the difficulty of analyzing data sufficiently,and the probability accuracy needs to be further improved.In view of the above situation,this paper deeply studied the theory,method,technology and application of accident safety control combined with accident prediction and intrusion monitoring.The results realize the whole process control of accident monitoring,model analysis and real-time alarm in safety supervision of construction site,which is valuable to improve the pertinence of safety measures,reduce the occurrence of safety accidents and improve the safety management of construction site.The main research achievements and innovations are as follows:(1)On the basis of accident classification and the relatd causes analysis,a fuzzy comprehensive evaluation method of construction accident was carried out by using the expert group decision method and the D-S evidence fusion theory.The method gives the main causes of construction safety accidents in large-scale projects,which overcomes the subjectivity of experts in traditional evaluation methods and enhances the credibility and objectivity of the evaluation results of safety accident causes.With analyzing on the correlations between main courses and unauthorized intursion,the unauthorized intrusion is an important attempted event,which provides a typical object for construction site safety control.(2)The monitoring technology of unauthorized intrusion in construction site based on Io T was put forward.Giving a construction safety control strategy based on intrusion monitoring and incident prediction,the overall framework of the unauthorized intrusion monitoring system on construction site was designed,which integrated with RFID,Bluetooth,mobile network,cloud server and other Io T technologies.For monitoring the unauthorized intrusion behaviors on the construction site,the intelligent hardhats,portable RFID triggers,and the Io T-based intrusion monitoring system was developed,which helps to improve the accuracy of detecting and position,alarming delay and failure to identify intruders,and it is more suitable for the job site environment and portable use.At the same time,the hidden danger information collected by the system provided basic data for accident prediction.(3)Based on the analysis of the evolution of "incident precursor-near-missaccident",a Copula-based Markov model was proposed to describe the uncertainty of near misses and the dependence of near misses at different time,thus establishing a construction site accident dynamic prediction model combining the historical data of near misses and the prior knowledge of experts.The prior distribution of accident occurrences can be obtained through the expert judgement.The Copula-based heuristic method was applied to quantify the stochastic correlation between accidents and near misses,which simplifies the assignment of conditional probabilities.Furthermore,historical observations of near misses were used for updating the probability of accident occurrence,and acceptance guidelines for determining accident-prone periods was given.This method combined with expert prior knowledge to modify the accident probability,so as to further improve the pertinence and accuracy of accident prediction.(4)A construction site safety inspection system based on the We Chat program was developed,and the method of recording near misses and hidden dangers on job site was given.The overall system framework of light applications based on the We Chat was designed to improve the development efficiency and realize multi-platform operations at one-end development.In view of the diversity of construction sites and the problem that the single user interface for data collection cannot meet the needs,a template-based system configuration algorithm based on the field application scenario was developed.Users can define the data collection template for specific inspection objects according to their own scenarios.This system cannot only be used as a supplementary means of on-site hidden danger inspection and safety accident control,but also increase the data source of accident prediction for the unauthorized intrusion monitoring system.