Deformation Prediction Of Deep Foundation Pit Based On GA-BP And Security Risk Assessment Research

With the deepening of urbanization,a large number of people have migrated to cities,underground space has been used more and more widely,and a large number of deep foundation pit projects have emerged.However,during the construction of a foundation pit project,there are long periods and many uncertain factors.Once an accident occurs,it is very likely to cause casualties and economic losses.Therefore,the deformation prediction and safety evaluation of the foundation pit is an extremely important task,and it runs through the entire construction process.There are many factors that affect the deformation and safety of the foundation pit.The change of any one factor will affect the final judgment result.Therefore,it is difficult to accurately predict the deformation of the foundation pit and evaluate its safety status.Real-time monitoring data can effectively reflect security status of the foundation pit.By constructing the relationship between monitoring data and time,predict and analyze the deformation of deep foundation pits.At the same time,the monitoring item data obtained from the deformation prediction is incorporated into the risk assessment system to evaluate the safety status of the deep foundation pit.Based on a deep foundation pit adjacent to a subway station in Zhengzhou City,this paper combines on-site monitoring data to carry out research on the deformation law,deformation prediction and safety risk assessment of the deep foundation pit,the main conclusions are as follows:1.Based on the project overview,surrounding environment,and geological conditions of the deep foundation pit,a monitoring plan for the deep foundation pit was formulated,and representative monitoring points in important monitoring items were analyzed to obtain changes in monitoring data over time curve,analyze the deformation law of each monitoring item.2.Discuss the theory and characteristics of BP neural network and genetic algorithm,and propose a way to optimize BP neural network by genetic algorithm to improve prediction accuracy.MATLAB programming determines the best BP neural network model and GA-BP model,respectively predicts the surface settlement of deep foundation pits,and compares the final prediction results with the monitoring results,and the average relative errors of the two prediction models are 2.6297% and1.7111%,the relative error increased by 0.9168%,indicating that the BP neural network optimized by genetic algorithm has higher prediction accuracy.3.By comparing and analyzing the advantages and disadvantages of various safety evaluation methods such as expert evaluation method,analytic hierarchy process,fuzzy comprehensive evaluation method,and the evaluation process,a combination of expert evaluation method and analytic hierarchy method is proposed to evaluate the safety risk of deep foundation pits.Take the cumulative change and rate of change of important monitoring items as safety risk assessment indicators,build a risk assessment system and security level,convert various monitoring data and cumulative change rate of a certain day into risk probability values,and combine the comprehensive evaluation base of risk loss values the risk level of the pit.