Research On Dynamic Prediction Method Of Deep Foundation Pit Deformation Based On Heuristic Algorithm Under Pile-brace Supporting System

The foundation pit engineering has the characteristics of ’large quantity,large depth and complex surrounding environment’.The accidents of foundation pit engineering occur frequently,which cause serious harm to the surrounding environment and workers of foundation pit.Therefore,during the excavation of the foundation pit,how to accurately predict the deformation,assess the risk of the whole body in real time has important engineering guiding significance.At the same time,combine the research results with the concept of intelligent engineering is also of great significance to promote the developing infrastructure supported by the state.This paper starts from the two most important indicators:horizontal and vertical displacement in the foundation pit monitoring project.Based on the practical project,a theoretical calculation model is established.A dynamic prediction and risk assessment model of foundation pit deformation refers to machine learning and heuristic algorithm is further constructed.Finally,a smart cloud platform for dynamic prediction and risk assessment of foundation pit engineering is developed.The dynamic prediction and risk prediction of the whole cycle during foundation pit construction are realized.The main research contents and results of this paper are as follows:(1)According to the Winkler model,the interaction between soil and pile is equivalent to spring,and the internal force calculation model of pile-brace supporting structure considering soil displacement is established.The hybrid method proposed in this paper is used to derive the model,and the horizontal displacement of supporting structure is solved.Through comparative analysis of results with traditional methods and practical cases,it is proved that the calculation model proposed in this paper is reasonable and accurate.(2)Considering the influence of precipitation factors when calculating the surface settlement outside the pit,combined with the basic analysis of seepage theory and the principle of effective stress,a reasonable simplification method of water surface curve is established.The additional stress calculation formula considering the precipitation curve is further established.The total surface settlement can be obtained using the layerwise summation method.Through the comparison with actual cases,the effectiveness of this method is proved.A parameter analysis is carried out to analyze the influence of load distribution width,permeability coefficient and precipitation depth on the results.(3)Based on the established horizontal and vertical deformation calculation model,a dynamic prediction model of deformation during excavation of foundation pit is established,combining with machine learning algorithm and heuristic algorithm.On this basis,the risk assessment system of deep foundation pit construction is established,the predicted value of deformation is finally transformed into an intuitive risk level,which is conducive to the construction side to grasp the health status of the whole project.(4)Encapsulate and integrate the above research contents,embed the established dynamic deformation prediction model and risk assessment model of deep foundation pit.A smart cloud platform for deep foundation pit by using the front and back-end development tools(VS code and IntelliJ)is developed.Functions inlcluding automatic monitoring,smart prediction,risk assessment and project management of the full-life cyclic deformation in the construction can be realized.(5)Based on the above research,relying on a foundation pit project in Wuhan.The practicability of smart cloud platform for deep foundation pit is verified,and the efficiency of monitoring and control level of foundation pit are improved.