Research And Application Of Construction Safety Warning Method Based On Convolutional Neural Network

Construction of large production activities dangerous and many unsafe factors,large investment,risk,and low content of construction technology,technical level is not high and the labor intensity is high,often makes the construction industry safety accidents is higher than other industries,at the same time using CNN to combining with the characteristics of construction industry and the actual engineering project,the construction of major hazards,accidents and risk of engineering construction information for advance control management,and advance the previous tracking adjustment,afterwards controls to adjust and accident prevention in advance,in order to find safety accidents and hazards information at any time,So as to nip construction safety accidents in the bud.This research can be devided as following:(1)In order to determine the theoretical boundary and key concepts of the research,the research topic and perspective are defined by sorting out the current research status of convolutional neural network at home and abroad,construction safety early warning of construction engineering projects,and safety field,thus laying a theoretical foundation for the research of the full text.(2)to the deficiency of traditional models for a targeted improvement,to avoid the negative effect on the early warning result,through the improvement of BP neural network to construct the construction safety accident early warning based on convolution neural network model,combined with the safety accident cause theory,considering the influence of technical factors on the construction safety accident,in person,content,environment and management,and other factors to establish the early warning index system of the introduction of technical factors,the construction safety accident early warning index system.(3)in order to cause accidents of building construction safety accident,the information such as hazard sources,disaster and risk analysis,and judgment influence construction safety risk of all of the information,research by feature extraction module,the maximum number of iterations and minimum rate determines the convolutional neural network model training network structure,the convolutional neural network performance test of the safety early warning,and trained network model,put forward can convergence speed and the promotion of the convolutional neural network early warning method of prediction accuracy.(4)Based on the safety index data of 26 projects in the Bohai Rim region,the feasibility of the convolutional neural network model in the practical application of construction engineering management is verified.Combined with the practical cases,the applicability of the convolutional neural network model in the construction safety warning of practical engineering projects is verified.On the basis of the existing research,this paper realizes the pre-warning of coupling among some construction site risk factors,which to a certain extent enricates the construction safety early warning research in China,helps to expand the application field of convolutional neural network,improve the construction safety management early warning research,and realize the pre-warning of construction safety accidents.