Analysis And Research On Risk Of Deep Foundation Pit Engineering Before The Construction

With the continuous development of city construction in China, all kinds of large underground engineering occurs, such as the basement of high-rise building, tunnels, municipal square, bridges base and subway stations and so on. The development of underground space promotes the development of the foundation pit engineering. Depth and scale of foundation pit excavation is increasing, and there are many uncertain factors. The design and construction is more difficult. Due to reasons of technology and management, many accidents occurred in the deep foundation pit engineering, such as settlement and craze of surrounding roads, cracking and tilt of neighbouring building, translocation and rupture of underground pipelines, collapsed of neighbouring building and casualties. They delay the construction period, lead to unnecessary economic losses and the adverse social impact. Therefore, in order to reduce the occurrence of deep foundation pit engineering accidents, it is necessary to study the risk of deep foundation pit engineering.In this paper, the research content is as follows:(1) The research situation and development process of foundation pit engineering risk are introduced, the development of the risk analysis method is gived a brief summary.(2) Foundation pit accident data were collected for statistical analysis. The foundation pit accident construction factors are regarded as the main reason, followed by the survey and design reasons. Therefore, reducing errors of survey and design can reduce the risk of foundation pit engineering.(3) Aimed at the risk of foundation pit supporting design scheme, a risk analysis method based on BP neural network is proposed. Combined with the mathematical theory of the probability theory, foundation pit engineering risk is analyzed quickly. Random sample is obtained by uniform experiment, and each random sample is calculated by the finite element software to obtain the corresponding output sample. On this basis, the BP neural network is used to establish the response relationship between random samples and output samples. Then, combined with the Monte Carlo simulation, the failure probability of each failure mode in the foundation pit engineering is calculated. Finally, foundation pit system failure probability is obtained through the method of conditional probability.(4) The deep foundation pit engineering in Kunming is as an example. The engineering risk is analyzed by the proposed methods. The soil shear strength index, soil thickness, depth of ground water level and surrounding load are selected as uncertain parameters. The main failure mode of the foundation pit engineering is pointed out. On this basis, the failure probability and reliability index of foundation pit engineering is calculated, and the risk is assessed.