利用背景噪声高阶频散曲线研究四川盆地结构

Study on the structure of Sichuan Basin by using high mode dispersion curves of ambient noise

张功恒

11749276

硕士

力学

陈晓非

2019-05-23

2019-05-31

哈尔滨工业大学

深圳

本文通过回顾背景噪声方法的研究现状，对从背景噪声数据中提取经验格林函数的理论基础进行了介绍，总结了从背景噪声数据中提取得到经验格林函数的数据处理方法，对处理流程进行了归纳，并通过模拟数据比较了从时间域计算互相关与从频率域计算互相关函数得到的频散能量图，发现从频率域计算互相关函数对于得到更高阶、更高信噪比的频散能量图更为有利；通过推导广义反透射系数方法计算频散曲线，做好正演计算，为反演阶段打下结实基础；频散曲线反演理论发展历史漫长，从最初的半波长法等经验方法发展到现在的各种仿生学算法层出不穷，对拟牛顿方法-BFGS校正-信赖域方法进行详细推导；矢量波数变换是最新提出的一种对面波信号进行频散能量成像的方法，该方法能够有效地从背景噪声互相关函数得到含有高阶频散能量的频散能量图，笔者在此基础上综合高分辨率Radon变换的原理推导了高分辨率矢量波数变换公式并进行了模拟数据试算，表明高分辨率矢量波数变换可以有效地提高频散能量图分辨率；由于背景噪声数据互相关得到的经验格林函数信噪比较差，对比了平滑去噪、小波阀值去噪、经验模态分解去噪、稀疏分解去噪在模拟数据上的效果，结果表明基于矢量波数变换的稀疏分解去噪效果最佳，能够有效地减少频散能量图中的噪声能量；将所述方法应用到四川盆地区域，通过四川盆地及其周边区域背景噪声数据的处理，得到了各研究子区域基阶频散能量强、分辨率好、连续性好、频带分布较宽、含有高阶频散能量的频散能量图。在得到频散能量图后利用反演算法得到了各研究子区域一维横波速度模型，验证了青藏高原地区在深度为20km-40km位置存在低速区这一地质结论，横波速度最低可达3.25km/s；四川盆地由于受沉积层影响，相速度在0.3Hz位置可达2.5km/s，反演结果显示沉积层厚度在5km-10km；四川盆地东侧及南侧在20km-40km深度存在低速区,但该区速度值高于青藏高原同深度速度值；云贵高原地区西部通过反演结果显示在60km-80km深度存在低速区。通过分析比较，阐述了基阶频散分叉及能量紊乱的产生原因是由于在浅部区域水平向速度存在差异。

In this paper,we review the research status of ambient noise method and the theoretical basis of extracting empirical Green's function from background noise data. The data processing method of extracting empirical Green's function from background noise data and the process flow is summarized. And through the simulation data, the dispersion energy map obtained by calculating the cross-correlation from the time domain and the cross-correlation function from the frequency domain is compared. It is found that the cross-correlation function calculated from the frequency domain could obtain the dispersion of higher order and higher signal-to-noise ratio. the dispersion curve is calculated by the generalized reflection and transmission coefficient method, which provides the basis for the forward modeling in the inversion phase; the theory of dispersion curve inversion has a long history, and it is developed from the initial half-wavelength method. Up to now, various bionic algorithms have emerged in an endless stream, and the quasi-Newton method-BFGS correction-trust region method is deduced in detail. Vector wavenumber transform is a newly proposed method for dispersive energy imaging of surface wave signals. A dispersion energy diagram containing high-order dispersion energy is obtained from the background noise cross-correlation function, we Based on the principle of high-resolution Radon transform, the high-resolution vector wavenumber transformation formula is derived and the simulation data is simulated. It shows that the high-resolution vector wavenumber transform can effectively improve the resolution of the dispersion energy map. The data obtained from cross-correlation has a poor signal-to-noise ratio, and compares the effects of smooth denoising, wavelet threshold denoising, empirical mode decomposition denoising, and sparse decomposition denoising on the simulated data. The results show that the sparse decomposition denoising effect is the best, and the noise energy in the dispersion energy map can be effectively reduced. The method is applied to the Sichuan Basin region, and the research results are obtained through the processing of ambient noise data in the Sichuan Basin and its surrounding areas. A dispersion energy map with high-order dispersion energy, strong resolution, better continuity, wide frequency band distribution, and high-order dispersion energy. After obtaining the dispersive energy map, the one-dimensional shear wave velocity model of each sub-region is obtained by using the inversion algorithm. The geological conclusion that the low-speed region exists in the depth of 20km-40km in the Tibet Plateau is verified. The shear wave velocity is as low as 3.25km/s; Due to the influence of sedimentary layers in the Sichuan Basin, the phase velocity can reach 2.5km/s at 0.3Hz, and the inversion results show that the thickness of the sedimentary layer is between 5km and 10km. The inversion results in the Yunnan-Guizhou Plateau show there is a low speed zone at a depth of 60-70 km on the west side of the upper mantle. Through analysis and comparison, the reason for the base-order dispersion bifurcation and energy disturbance is explained by the difference in horizontal velocity in the shallow region.

频散曲线 矢量波数变换 经验格林函数去噪 背景噪声

High mode dispersion curve VWTM Empirical Green's function denoising Ambient noise

中文

联合培养

南方科技大学

张功恒. 利用背景噪声高阶频散曲线研究四川盆地结构[D]. 深圳. 哈尔滨工业大学,2019.