Research And Application Of Swarm Intelligent Optimization Method In Rayleigh Wave Dispersion Curve Inversion

In the research and application of near-surface investigation,surface wave exploration is an effective method to detect shallow subsurface geological structures and physical parameters,widely used in urban underground space exploration.The Rayleigh wave field orthorectified simulation is a meaningful way to study the Rayleigh wave propagation law,which reveals the absorption and attenuation characteristics of Rayleigh waves in the seismic records corresponding to different geological models;the Rayleigh wave dispersion curve implies the information of the physical parameters of the subsurface medium.The cross-wave velocity structure can visually reflect the anomalous structural changes in the subsurface,which is essential for early warning of geological hazards.Therefore,it is crucial to study the inversion method of the Rayleigh wave dispersion curve with strong adaptability and high inversion accuracy to enhance the processing and interpretation of surface wave exploration,which is one of the urgent problems to be solved nowadays.This paper systematically investigates the forward simulation of the Rayleigh wave field and Rayleigh wave dispersion curve in an elastic medium,viscoelastic medium and anisotropic medium;compares and analyzes seven calculation methods of Rayleigh wave dispersion energy spectrum,evaluates the imaging resolution and dispersion curve extraction accuracy of different imaging methods in the frequency-velocity domain;constructs dispersion curve inversion method of the Rayleigh wave based on intelligent swarm optimization algorithm to improve the accuracy of surface wave This paper presents a multi-axis complex algorithm based on the multiaxial imaging method.In this paper,we propose a numerical simulation of Rayleigh wave field orthorectification based on the absorption boundary of multi-axis complex frequency shift approximation perfectly matched layer(MCFS-NPML),and systematically discuss and compare the Rayleigh wave propagation law and dispersion curve trends in an elastic medium,viscoelastic medium and anisotropic medium under finite boundary,uniform half-space,three-layer horizontal laminar medium model and cavity-containing stratigraphic model,respectively.Since the absorption boundary can lead to false reflected waves and Rayleigh wave amplitude energy decay,it is verified that MCFS-NPML can effectively improve the orthorectification algorithm’s computational efficiency and simulation effect.Based on the Rayleigh wave field orthorectified simulation,different stratigraphic velocities’ dispersion curve variation characteristics are analyzed based on the elastic medium and viscoelastic medium layered model.The vector wavenumber transformation method is used to calculate the Rayleigh wave dispersion energy spectrum,which improves the imaging accuracy of the Rayleigh wave dispersion energy spectrum and realizes the utilization of Rayleigh wave multimodal dispersion curves.The Rayleigh wave multimodal dispersion curve change characteristics cannot be observed directly from seismic data,so it is necessary to convert the seismic signal from the time domain to the frequency-velocity domain by data processing method to extract the Rayleigh wave dispersion curve,which has significant practical value as the basis of subsequent dispersion curve inversion.Therefore,in this paper,we conduct an in-depth study of the calculation methods of the Rayleigh wave dispersion energy spectrum.In the three-layer elastic medium horizontal laminar model,we study seven calculation methods of Rayleigh wave dispersion energy spectrum,such as - transform,- transform,SFK transform,and vector wave number transform,in the frequency-velocity domain imaging resolution and applicability.The results of model orthorectified data processing show that the Rayleigh wave dispersive energy spectrum calculation method of vector wavenumber transformation method can achieve high precision dispersive energy spectrum imaging and effectively improve the extraction results of the Rayleigh wave dispersion curve.Therefore,the vector wavenumber transformation method is more applicable to calculate the Rayleigh wave dispersion energy spectrum,which can better reflect the near-surface structure and physical parameters and provide technical support for the subsequent Rayleigh wave data processing.This paper proposes two methods of Rayleigh wave dispersion curve inversion based on the adaptive step-improved Firefly algorithm and the improved Ocean Predator(ACMPA)algorithm.By performing adaptive step improvement on the firefly algorithm,the step function that best fits the algorithm is selected according to the transformation trend of different step functions with the number of iterations.This can effectively improve the convergence speed and algorithm accuracy.On this basis,the step function of the marine predator algorithm is improved;the quality of the initialized population position is improved by the chaotic initialization method;the search capability is enhanced based on the elite competition method to avoid the algorithm falling into the local extreme value point,and the search range is dynamically adjusted to improve the convergence speed of the inversion from both local and global search for the best.The improved adaptive step size,chaotic initialization and elite competition mechanism based on the We propose an improved Ocean Predator(ACMPA)algorithm based on adaptive step size,chaotic initialization and elite competition mechanism.The multi-peak mathematical function model is tested and analyzed by PSO(Particle Swarm Optimization)algorithm,FA(Firefly Algorithm)algorithm,damped least squares method,MPA(Marine Predators Algorithm)algorithm,improved firefly algorithm and improved marine predator optimization algorithm.The convergence and stability of the improved Marine Predators Algorithm(ACMPA)are proved to be excellent,and the convergence accuracy and convergence range are better than other optimization algorithms.By constructing the inverse problem corresponding to the Rayleigh wave dispersion curve and applying the inversion to the forward simulated seismic records and the measured seismic data based on the ACMPA algorithm,it is demonstrated that the method for inversion of Rayleigh wave dispersion curves on the ACMPA optimization algorithm has the advantages of strong adaptability and high inversion accuracy.