Study On Subway Construction Safety Evaluation Of BP Neural Network Model

With the rapid development of urban economy, the urban rail transit immediately increasing rapidly. Due to its green pollution 、 transportation speed 、 saving land 、 easing urban transport,Subway as the main underground operating mode of urban rail is pressed ahead with a number of cities. Currently there are 24 cities subway project put into operation, since 2013 the urban rail transit NDRC approval authority delegated to the provincial government, our country entered into the peak of subway construction. Subway construction in China in 2020 is expected to reach more than 45 cities. However, in the process of development of ultra-fast subway, the frequent safety accidents are worth pondering alert point. In our country, underground development is just only about 40 years of construction experience, far apart developed countries. Once construction accidents in the event, it will cause serious casualties and huge economic losses to the community property. In view of the above, subway construction safety issues are imminent, therefore subway construction safety evaluation is very important.The purpose of this paper is to establish a safety assessment prediction model, through a comprehensive analysis model evaluate the subway project safety levels to predict the types of accidents most likely to occur, so that the construction unit can take corresponding measures for this result. Take the theory of identification of hazards on the basis to confirm subway construction safety factor in order to establish safety factor evaluation index; count nearly four years of construction accident types, comprehensive analyze the accidents and determine the four-prone as security incident evaluation index. Both together constitute the subway construction safety evaluation system, while the safety factor as BP neural network model training input factor, and accident factor as model output factor. By constantly training and learning the model, finding the nonlinear law hidden between the safety factor evaluation index and safety accident evaluation index and the eventual establishment of suitable construction safety evaluation metro network model is completed.. In this paper, it offers a new method for subway construction safety evaluation, namely subway construction safety evaluation of BP neural network prediction model. Finally the BP neural network is predicted about the probability of occurrence accidents of Cheung Tsing Bridge Station ~ Hunnan Avenue station interval of Shenyang Metro Line 10. According to this result, the construction units can take safety management measures. It has important theoretical significance for the comprehensive safety assessment of subway project.Study on safety evaluation of subway construction prediction can be found in advance of major hazard based on the results. Safety evaluation of subway can improve the subway construction safety management and control early warning capabilities, the maximum possible to avoid or reduce our subway construction process safety incidents. At the same time, it provides security for China to promote urban rail transit construction process, and it is of great important theoretical significance for the future development of comprehensive safety evaluation metro project.