Font Size: a A A

Displacement Back Analysis Method For Physical And Mechanical Parameters Of Rock Mass Based On Immune And Gaussian Process Coupling Algorithm

Posted on:2019-09-02Degree:MasterType:Thesis
Country:ChinaCandidate:Z H WangFull Text:PDF
GTID:2370330545965662Subject:Civil engineering
Abstract/Summary:PDF Full Text Request
Geotechnical engineering is a comprehensive discipline that is highly unpredictable and it is difficult to determine the parameters of geotechnical engineering material properties.This is also an obstacle to the study of various phenomena and behaviors in this field.After long-term development,people proposed an inversion analysis method to solve this problem and achieved good results.The back analysis of mechanical parameters of tunnel engineering has always been an important issue in rock mechanics.Based on the analysis of different computer theories,the displacement back analysis method can be further subdivided into numerical method and analytical method.The numerical method has a wider range of applications.Typical numerical methods include difference method,discrete element method,boundary element method and finite element method.The optimized inversion method is the most widely used.In recent years,the displacement back analysis method based on intelligent calculation has been rapidly developed to solve the problem of three-dimensional displacement back analysis of tunnels and underground engineering.This paper innovatively uses Immune Algorithm(IA)and Gaussian Process(GP)coupling(immuno-Gaussian process algorithm)to invert rock physical and mechanical parameters.By introducing a combination kernel function to improve the generalization performance of the Gaussian process regression,engineering verification was carried out by predicting the deformation time series of the new landslide in Wolongsi.This paper systematically analyzes the results of numerical experiments based on the immune Gaussian process algorithm,and uses this as a sample for network training.It fully considers the displacements of tunnel monitoring points at different distances from the monitoring section and the face surface,and is described by network training.The Gaussian Process Regression(GPR)intelligent model of nonlinear mapping between surrounding rock displacement and physical and mechanical parameters of rock mass is used to input the measured surrounding rock displacement into this model,and the immune algorithm is used to search the range of parameters to be inverted.The internal automatic search can make the GPR predict the physical and mechanical parameter combination of the displacement and the measured displacement closest to the actual displacement,and complete the inversion of physical and mechanical parameters of the surrounding rock.Then the inverse rock mechanics parameters are obtained by inversion calculation and used as the input parameters of the intelligent model,so that the displacement of the measuring point from the face to the different monitoring sections and different positions can be obtained.The FLAC3D software was used to verify the results.The results show that the immune Gaussian process algorithm intelligent model has a high degree of accuracy.The actual application results based on tunnel engineering show that the error in the three excavation steps does not exceed 20%.When there are more than three excavation steps,there is a large error between the measured value and the displacement prediction value.Therefore,at the time of excavation to step 4,the displacement should be recalculated with the current measured displacement,and then the estimated displacement value of no more than 3 excavation steps should be obtained.This model can also be used to predict the stability limit value of the surrounding rock during construction,ensure the safety of the construction,and can be used to evaluate the quality of the construction and reduce the construction cost to a certain extent.Based on genetic support vector regression algorithm,genetic Gaussian process regression algorithm,immune support vector regression algorithm and immune Gaussian process regression algorithm,the inversion of physical and mechanical parameters of rock mass and the prediction of advanced displacement are performed for the Beikou tunnel project.The tunnel construction and design provide important reference and basis.
Keywords/Search Tags:Displacement back analysis, Immune Algorithm, Gaussian Process, Combined kernel function, Parameter inversion
PDF Full Text Request
Related items