Abstract: Deep foundation pit engineering widely exists in the process of urban construction. It is an important part of disaster prevention and mitigation to analyze the monitoring data and study the environmental effects in the process of deep foundation pit construction. It is of great significance to analyze and predict the changes of the surrounding environment during the construction of deep foundation pit by using the monitoring data with high quality and combining the artificial intelligence technology. Relying on the pipe jacking project, this paper analyzes the impact of the excavation of the foundation pit of the starting shaft on the surrounding environment. The important monitoring parameters are processed into a time series, and advanced cyclic neural network (LSTM) is employed to achieve the advance prediction of monitoring parameters. The results show that there is a linkage law among the monitoring variables, but the linkage of the monitoring variables is relatively weak due to the lack of monitoring data or the impact of construction accidents, and the prediction of the cyclic neural network is not ideal. These works are of forward-looking significance for improving the early warning theory of foundation pit.

Key words: Deep foundation pit, Environmental effects, monitor, Long and short-term memory network, artificial intelligence, Early warning