The Prediction Model And Safety Assessment Analysis Of Subway Structural Deformation

As an important part of urban transportation, subway can make a significant contribution to urban transport, at the same time, which may also be lead to accidents resulting in casualties and huge economic losses to the city due to the structural deformation and other factors. The structural deformation of subway is one of the keys to the structural safety problems of subway, but also a very important indicator about subway security operating environment, thus how to ensuring the safe operation of subway structure deformation is a problem which is worthy of further studying. Subway operates for a long time, but the inspection and maintenance time is short, and the factor leading to the structural deformation of subway is very complex, therefore how to quickly and effectively achieving the safety assessment about the structural deformation of subway has a very important significance for the safety of subway operation.Effective decisions are based on data and analysis. This paper conducted research on the prediction models and safety assessment methods of subway structural deformation with measured deformation data. Specific research divides into the following areas:(1)The measured data of subway structural deformation was analyzed, and then using the characteristics which subway structural deformation has the periodic changes over time, established multiple regression analysis predictive models, and combining with the subway project examples compared the prediction accuracy of each regression model, and ultimately determined the regression analysis optimization model which appropriates for deformation prediction of subway construction.(2) The author proposed BP-regression analysis fusion prediction model, based on studying regression analysis predictive model and BP neural network predictive model. By two engineering examples subway project A and B, compared and analyzed prediction accuracy of the three models about regression analysis predictive model, BP neural network prediction model and BP-regression analysis fusion prediction model. For subway project A, the prediction accuracy of regression analysis model is ±0.229mm, and the prediction accuracy of BP neural network model is ±0.123mm, while the prediction accuracy of BP-regression analysis fusion model prediction is ±0.088mm, which improves by 28.5% compared to the BP neural network, and improves 46.3% compared to the regression analysis model. For subway project B, the prediction accuracy of regression analysis model is ±0.588mm, and the prediction accuracy of BP neural network model is ±0.355mm, while the prediction accuracy of BP-regression analysis fusion model prediction is ±0.231mm, which improves by 34.9% compared to the BP neural network, and improves 60.7% compared to the regression analysis model.(3) Through in-depth analyzed the factors of affecting the subway structural deformation, the author established the security evaluation index system about structural deformation of subway. The author told the modeling ideas of hierarchy-fuzzy comprehensive evaluation model. Several key issues in the evaluation model, such as security evaluation grade series, each evaluation criteria, the determination method of each evaluation index weights and the subjection degree vector, subjection function, and quantifying the final results of the evaluation, the author put forward the corresponding solutions, thus established hierarchy-fuzzy comprehensive evaluation model which based on the subway structural deformation security.(4) The author used the subway structural deformation security evaluation model established in this paper to assess the security situation about structural deformation of a subway engineering, and made some recommendations, which verified the practicality and effectiveness of the assessment model. The authors developed the software about subway structural deformation security state assessment.