Font Size: a A A

Back Analysis Of Soil Parameters Of Deep Foundation Pit Based On LSSVM And GA Algorithm And Deformation Prediction Of Retaining Piles

Posted on:2020-09-27Degree:MasterType:Thesis
Country:ChinaCandidate:X MaoFull Text:PDF
GTID:2392330578464988Subject:Architecture and civil engineering
Abstract/Summary:PDF Full Text Request
In geotechnical engineering,due to the anisotropy of rock and soil,the determination of geotechnical parameters is extremely difficult,which makes the calculation results of the model different from the actual ones.In recent years,with the rapid development of on-site monitoring and inversion analysis technology,the whole process of foundation pit construction is monitored with a reasonable monitoring plan.According to the construction conditions and monitoring results,comprehensive and reliable construction information is obtained.The deformation of the supporting structure is predicted and the subsequent construction plan is dynamically adjusted.Based on the crater reconstruction project in Nanxi District of Yibin City,based on the genetic algorithm(GA)optimization,the least square support vector machine(LSSVM)model is used to inversely analyze the elastic modulus of the foundation pit and use it.The feasibility of inversion parameters to predict the deformation of subsequent operating conditions is discussed.First,the monitoring data of the foundation pits are sorted and analyzed.Secondly,the representative monitoring points are selected,and the MIDAS GTS finite element software is used to simulate the working conditions of the foundation pit construction,and the simulation values and monitoring values are compared and analyzed.Then,using the uniform experimental design and construction parameter combination to bring into the established model to calculate the test sample,based on the test sample and GA algorithm to establish the LSSVM model,based on the model to reflect the elastic model of the foundation pit soil layer through the monitoring value the amount.Finally,the elastic modulus obtained by the inversion is brought into the numerical model for verification,and the parameters verified as reliable are predicted for the subsequent working conditions.The main contents of this paper are as follows:(1)After introducing the basic outline of the project,the axial force of the anchor cable,the horizontal displacement of the retaining pile and the bending moment of the pile body were analyzed.Based on the enumeration of the factors causing the axial loss of the anchor cable,the anchor cable axial force of the project is analyzed based on the monitoring information.Based on the monitoring data,the overall state of the pile deformation is analyzed,and the maximum horizontal displacement and excavation are counted.The ratio of depth is calculated.The bending moment of the pile body is calculated based on the monitoring value of the steel bar stress,and the variation process of the bending moment during the construction process is analyzed.(2)selection and parameter monitoring points provided in the exploration report representative,binding MIDAS GTS finite element software foundation construction process is simulated,and the simulation results of the monitoring results were analyzed.(3)A brief introduction to the theory of inversion analysis is made.The elastic modulus E of the soil with large influence on the horizontal displacement of the supporting structure and the difficulty of determining the size is selected as the parameter to be analyzed.The monitoring is convenient,the data is comprehensive and the precision is better.The horizontal displacement of the high support structure is used as the measured data in the inversion analysis.By uniform design embodiment,design parameters of a plurality of sets of soil composition,Chapter excavation using the model to calculate the resulting model of learning samples instead of MIDAS GTS with LSSVM calculation model being analyzed The nonlinear mapping relationship between the elastic modulus of the soil and the horizontal displacement of the deep body of the pile is used to search and establish the LSSVM benchmark model by searching the optimal regularization parameter ? and kernel parameter ? of the LSSVM model.Thus,for the horizontal displacement of the pile body under any working condition,the elastic modulus of each soil layer under the working condition can be obtained by using this model.(4)The results of the inversion analysis of each working condition are brought into the model for calculation,and compared with the measured data of each working condition in the actual construction process,and the results of the inversion analysis are verified to be reliable.Finally,the verification is used as a reliable parameter for the subsequent working conditions.prediction.
Keywords/Search Tags:Deep foundation pit engineering, foundation pit deformation prediction, MIDAS GTS, least squares support vector machine, genetic algorithm
PDF Full Text Request
Related items