Displacement Back Analysis Of Deep Foundation Pit Based On Support Vector Regression And PSO

The process of construction of deep foundation pit engineering will produce displacement, heave, the surrounding ground surface subsidence deformation of retaining structure. There are many factors that affect the deformation of deep foundation pit, which is a complex nonlinear problem. According to the requirement of information construction, monitoring is essential in the process of deep foundation pit construction. Based on existing monitoring data, the deformation of foundation pit can be predicted, which can guide the construction of foundation pit monitoring. In the process of foundation pit construction, the supporting structure and soil parameters and other factors affect the displacement of foundation pit. Soil parameters are one of the important factors affecting the deformation. How to obtain a parameter that can not only reflect the characteristics of soil, but also can reflect the influence of support construction, and then use the existing monitoring data to predict the future deformation, which has the important significance of reality.Based on the monitoring data of deep foundation pit engineering in Xinzhou Weinan real estate F plot, this paper researchs the application of support vector regression machine(?-SVR) and particle swarm optimization algorithm(PSO) in foundation pit displacement back analysis and prediction. The monitoring data of each measurement point is analyzed, and the control significance of the whole foundation pit deformation is analyzed. The test parameters of 25 groups of soil mass are designed by the method of orthogonal experiment, and the positive analysis was carried out by FLAC3 D software. Based on the simulation results, the support vector machine regression model is established to reflect the mapping relationship between the parameters of the soil and the deformation simulation. Based on the support vector regression model, the particle swarm optimization algorithm is used to search the optimal parameters with the minimum of the measured deformation error as the objective function. The FLAC3 D model calculation is carried out by using the parameters obtained by back analysis.,predict the following conditions of the foundation pit supporting structure deformation.The contents and results of this paper are as follows:(1) During the construction of the foundation pit, the relative monitoring work is more comprehensive. The horizontal displacement of the pile and the surrounding ground settlement are all within the early warning value. The horizontal displacement of CX9 pile body on the west side of foundation pit is the largest, and the horizontal displacement of the pile under each working condition can reflect the influence of construction. It is reasonable and important to analyze and predict the displacement monitoring data of the measuring point, which is important for the whole foundation pit deformation control.(2) The construction process of each working condition is simulated by FLAC3 D software. The simulation results can reflect the deformation characteristics of the foundation pit. With 25 sets of soil parameters which are determined by the method of orthogonal uniform design, FLAC3 D model is used to calculate the horizontal displacement of pile body,so as to establish a support vector regression model.(3) According to the field monitoring data, based on the support vector regression model, the elastic modulus of the soil mass is searched with the minimum measured displacement error as the objective. The inversion parameters are substituted into the model of FLAC3 D calculates the value of displacement and and compared with measured displacement, back analysis of return calculated value and measured the displacement error is small, indicating that the back analysis of soil parameters can reflect the characteristics of the soil, and can reflect the supporting construction effect.(4)The inversion parameters are substituted into the model of FLAC3 D to calculation of the following conditions. The predicted values of deformation are obtained, and the predicted values are in good agreement with the measured values. It can be considered that the back analysis parameter has a certain value for the prediction of the following condition.