Surface Wave Dispersion Analysis And Vector-processing Methods For Multi-component Seismic Data

Recently,vector-processing and vector-interpretation of multi-component seismic data are developed with the development of multi-wave seismic exploration technique.The velocities,amplitudes,and frequencies of P-and PS-waves are obtained from the seismic vector fields.In addition,the polarizations of the seismic wave and the differences of the time,amplitudes,and phases about different waves can be obtained from the seismic vector fields.In engineer prospecting,the vertical component is used alone for traditional surface-wave methods.The incomplete and inaccurate dispersion curves are obtained from it,which will result in incorrect velocity structures in shallow layers.More information about underground medium can be obtained from multicomponent surface-wave vector fields.It is questionable how to use surface waves in different components.To obtain a reliable information about the seismic vector fields,the vector characteristics need to be kept when multi-component seismic data are pre-processed.However,the traditional methods of denoising and separation of the wavefields are usually processing each component separately,which damages the vector characteristics.So,they are not applicable for multi-component seismic data.There is urgent need to design appreciable denoising methods which can also keep vector characteristics of multi-component wavefields.The study on vector-processing methods for multi-compoent seismic data is the center of this thesis.The vector-processing methods for multi-component surface waves and the methods of multi-component seismic data denoising are discussed in the thesis.The key points are as follows.The first work is the theoretical study of multi-component surface waves.The calculation of surface-wave apparent velocities of are developed to multi-component surface waves.The dispersion characteristics of multi-component surface waves acquired by multi-component geophones and activated by multi-component sources are discussed.The cause of the mode jumping phenomenon is explained.In addition,the relationship of the multi-mode ellipticity and the parameters of shallow layers is discussed.The second work is to design the vector-processing methods for multi-component surface waves.The method of extraction of Rayleigh-wave dispersion curves based on complex-vector seismic data is proposed.Based on different dispersion characteristics of the radial and vertical components,the dispersion images of the complex–vector seismic data show good performance against interferences and mode misidentification,which results in more complete and accurate dispersion curves compared to the traditional vertical-component Rayleigh-wave method.The asymmetric surface-wave dispersive energy of complex-vector seismic data is utilized to separate the fundamental mode and the higher mode of Rayleigh waves by the polarization filtering method.The third work is to design the methods of multi-component seismic data denoising based on sparse representation and quaternion.A method for surface-wave extraction is proposed based on the morphological differences between surface waves and reflections to obtain high-fidelity surface waves in the X component.The quaternion dictionary learning is introduced to seismic vector field processing from the picture processing.The dictionary trained by the K-QSVD algorithm is applicable to threecomponent seismic data denoising.An improved dictionary learning method dedicated to seismic vector data is proposed,which keeps the vector characteristics,improves the signal-to-noise ratio,and decreases the calculation time of dictionary learning.