Chengdu Metro Jinjiang Station Deformation Prediction Research

Deformation prediction is very important for foundation pit excavation, can predict the deformation of foundation pit during excavation, and judge whether there is a potential safety problems, reduce the foundation pit accident. However, geotechnical material is complicated, lead to deformation unpredictable. In this paper, Combining with the finite element analysis and the BP neural network, and in south extension line of chengdu metro line 1 jinjiang station foundation pit engineering study on deformation prediction analysis as the research object. This article main research content is as follows:(1) for deformation prediction at home and abroad, the research status of the comprehensive analysis of conventional analysis method were introduced in detail, the basic principle of finite element analysis, and system analysis, the main development course and existing main problems, summarizes the main research direction.(2) the south extension line of subway line 1 jinjiang station foundation pit engineering general situation of the project, construction conditions and made a comprehensive narrative foundation pit monitoring, engineering characteristics, the deformation characteristics of foundation pit are summarized, providing the analysis of the deformation prediction with the experimental data.(3) based on the construction condition, numerical simulation was carried out on the jinjiang station foundation pit, the foundation pit lateral deformation of the ground subsidence, uplifts, characteristics and deformation of the retaining structure data, and analyzed in the foundation pit testing data, Shows that numerical simulation can be simple forecast of foundation pit deformation law.(4) Using the retaining piles with the field measured values and the numerical simulation of the difference between the value as input factors, respectively, using BP neural network to forecast the deformation of foundation pit retaining structure level, contrast analysis and numerical simulation calculation and BP neural network calculation value of prediction effect, showed that using difference has the highest reliability for deformation prediction, the error of the numerical simulation is the largest.