| Rayleigh wave is a kind of seismic wave formed by the interference between the Primary wave and the vertical component of the Secondary wave along the free interface of the medium;Rayleigh wave has strong energy,slow attenuation,high signal-to-noise ratio,and dispersion characteristics in layered medium,so Rayleigh wave is widely used to obtain the S-wave velocity structure of the subsurface medium.In near-surface geophysical exploration and research,Rayleigh wave exploration has been widely used with its characteristics of non-destructive,high efficiency,and high accuracy.The inversion of Rayleigh wave dispersion curves to obtain the S-wave velocity profile of near-surface layered medium is a committed step of Rayleigh wave exploration.The current inversion methods are mainly divided into local linear optimization methods and global nonlinear optimization methods.The accuracy of the traditional local linear inversion depends on the initial model and requires the calculation of partial derivatives,while the global nonlinear inversion avoids these problems.The conventional Rayleigh wave dispersion curve inversion is carried out for the fundamental mode dispersion curve of the Rayleigh wave,but if the subsurface structure is relatively complex,there are obvious velocity inhomogeneities,and there is a low-velocity layer or high-velocity layer in the half-space,the higher modes dispersion curves have greater penetration depth than the fundamental mode,and they are also more sensitive to the changes of stratigraphic parameters than the fundamental mode,so the effect of using only the fundamental mode dispersion curve inversion is often unsatisfactory,and the introduction of higher modes will improve the accuracy of the inversion results.However,the traditional multi-mode dispersion curve inversion simply assigns subjective weights to the objective functions and finally fits them together,ignoring the contradictions and conflicts among the objective functions;given this,this thesis uses a global nonlinear multi-objective optimization algorithm to invert the multi-mode dispersion curves to overcome the above problems.In this thesis,we first summarize the current research status of Rayleigh wave dispersion curve inversion through literature research and propose a multi-mode Rayleigh wave dispersion curve multi-objective optimization inversion strategy based on the existing research to address the shortcomings in the inversion algorithm and the joint inversion of multi-mode dispersion curves.The main concepts of multi-objective optimization theory are introduced;the third generation of fast Non-dominated Sorting Genetic Algorithms Ⅲ(NSGA-Ⅲ)is used as the optimization algorithm for inversion through the screening of test functions,and the basic concepts of the algorithm and the implementation process are briefly introduced.Then,the advantages and effectiveness of this inversion strategy are verified by the inversion of a large number of different theoretical models and comparison with the inversion results of other algorithms;then the excellent global search ability of the NSGA-Ⅲ algorithm in the absence of a priori information is verified by changing the search range of inversion.The theoretical model is then added with different degrees of random noise to test the validity of this study from a close to the actual situation.Finally,the inversion of the measured data is performed to check the feasibility and practicality of the application of this study in practical exploration and research by example.The study of this thesis shows that:(1)The NSGA-Ⅲ algorithm has superior performance in multi-objective problems,so the algorithm is applied to the inversion of multi-mode dispersion curves;the feasibility of the algorithm in the inversion is tested by a large number of theoretical model trial calculations.(2)The effect of using a multi-objective optimization algorithm to invert multi-mode dispersion curves is better than that of using a traditional single-objective optimization algorithm.(3)The NSGA-Ⅲ algorithm has a strong global search capability and even given a wide search range,this study can effectively invert the model and successfully characterize the geological model;it is also learned that the setting of the search range should consider the complexity of the model with the minimum and maximum phase velocity of the dispersion curve,and for more complex models,too large a search range will lead to low computational efficiency.(4)The difficulty of the inversion can be judged by analyzing the distribution of the Pareto front,and the quality and accuracy of its inversion will be affected to a certain extent as the difficulty of the inversion changes.(5)Through the inversion of a large number of theoretical models containing different degrees of random noise and measured data can be learned,the use of multiobjective optimization algorithm inversion of multi-mode dispersion curve research has practical and certain development potential.The main innovations of this thesis include the use of a multi-objective optimization algorithm to invert the Rayleigh wave multi-mode dispersion curves,which improves the inversion accuracy and avoids a series of problems that occur when using a traditional single-objective optimization algorithm to invert multi-mode dispersion curves;the relationship between Pareto front and inversion results is discussed.This study provides a new research idea for the joint inversion of Rayleigh wave multi-mode dispersion curves. |
PDF Full Text Request