地震层析成像原理与方法

需积分: 50 9 浏览量更新于2024-07-23 3 收藏 7.42MB PDF 举报

地震层析成像原理地震层析成像原理是指通过地震波传播的速度和方向来重建地球内部结构的技术。该技术广泛应用于石油勘探、矿产资源探测、环境监测和灾害监测等领域。该技术的核心是通过地震波的传播速度和方向来确定地球内部的结构和物质组成。根据给定的课程笔记，我们可以总结出以下几个重要的知识点： 1. 地震层析成像的基本概念地震层析成像是指通过地震波传播的速度和方向来重建地球内部结构的技术。该技术的基本概念包括forward problem、inverse problem、model parameterization、single-scale tomography和multi-scale tomography等。 2. 地震波传播速度的推断地震波传播速度的推断是地震层析成像的核心步骤。通过对地震波传播速度的分析，可以推断地球内部的结构和物质组成。地震波传播速度的推断方法包括直接推断和间接推断两种。 3. 地震速度的行为特征地震速度的行为特征是指地震波传播速度在不同岩石中的变化规律。地震速度的行为特征包括泛指趋势、实验测量和理论模型等。 4. 反射旅行时间层析成像反射旅行时间层析成像是一种常用的地震层析成像方法。该方法通过对反射波的旅行时间进行分析，来重建地球内部的结构。 5.utenberg-Wiechert旅行时间反演 Herglotz-Wiechert旅行时间反演是一种常用的地震波传播速度推断方法。该方法通过对地震波传播速度的分析，来确定地球内部的结构和物质组成。 6. 线性最小二乘逆推断线性最小二乘逆推断是一种常用的地震层析成像方法。该方法通过对地震波传播速度的分析，来确定地球内部的结构和物质组成。 7. 矩阵分解矩阵分解是一种常用的线性代数方法。该方法广泛应用于地震层析成像、信号处理和数据分析等领域。地震层析成像原理是指通过地震波传播的速度和方向来重建地球内部结构的技术。该技术广泛应用于石油勘探、矿产资源探测、环境监测和灾害监测等领域。

H. Zhou: 05/23/08 Seismic Velocity Analysis

steps. Firstly, it defines a set of sub-models with different cell sizes, and each sub-model

covers the whole model region. For the 1-D traveltime problem, Figure 1-6 shows five MST

sub-models each has layers that are multiples of the layers of the SST model. Hence, the

first order sub-model is the same as the SST model shown in Figure 1-5c, and the fifth

order sub-model which has only one variable, the average value of the whole model region.

The sub-models allow a decomposition of velocity anomalies into components of different

cell sizes.

As the second step, the MST determines all sub-model values simultaneously rather

than progressively. Progressive multi-scale inversion has been proposed in waveform

studies (e.g., Bunks et al., 1995) and data integration (e.g., Yoon et al. 2001). Many

previous tomographic studies have also attempted with multiple sets of cell sizes

progressively (Fukao et al. 1992; Bijwaard et al., 1999). In contrast, a simultaneous

determination of the parameters of all sub-models in the MST will minimize the spread of

model values between different model cells. Each model location belongs to all sub-models,

whiles at the same location the consistency between the data contributions, such as the

apparent slowness of different rays, may vary over different sub-models. The sub-model

with a higher level of consistency between data contributions will gain more from the

inversion. Panels (b) to (f) in Figure 1-6 illustrate possible solutions of the MST. Due to the

weight of their sizes, larger cells would take more model values when they match well with

the geometry of the true model anomalies. The sum of velocity values of all sub-models

should approximate the true velocity profile at each depth.

As the third and final step, the final MST model is a superposition of all sub-model

values. Hence a model variable in Equation (1-10) becomes

∑

(k)

, (1-12)

where K is the number of sub-models and

(k)

w is a weighting factor which can be defined a

priori. The unbiasedness condition in statistics requires that the sum of all weighting

factors equals to one, i.e.

∑

(k)

= 1, (1-13)

H. Zhou: 05/23/08 Seismic Velocity Analysis

kernels of the SST, the computation overhead of the MST is only marginally greater than

that of the SST.

Partition of the MST sub-models.

It is easy to partition the sub-models of the MST with cell sizes that are multiples of

the SST cells. Hence the differential kernels of the MST can be derived from that of the

SST. Taking a 2-D example, suppose that the cell size of the SST model is 1x1, then the

first order MST sub-model may use exactly the SST model dimension, the second order

MST sub-model may have cell size 2x2, the third order sub-model may have cell size 3x3,

and so on. With such a partition of the MST sub-models in 2-D, if the SST has N

cells, the

total number of MST cells would be N

+ N

+ N/3

+ N/4

+ … + 1 < N

x 165%. This

means that the increase of the number of inversion variables from the SST to such a 2-D

MST partition is less than 65%. For a 3-D case of similar model partition, the increase in

the number of inversion variables from SST to MST is only 20.2%. In other words, the

increase of model variables from the SST to MST is very moderate for both 2-D and 3-D

applications.

Just like the conventional tomography, there are infinite ways to partition the MST

model. Another way, for instance, is to partition the sub-models using cell sizes that are

prime-number multiples of the cell size of the SST. In other words, if the SST cell size is 1,

the MST cell sizes will be 1, 2, 3, 5, 7, 11, 13, and 17 and so on, and the largest MST cell

would be the whole model space.

Nevertheless, the most important consideration in partitioning the MST sub-models

is to have cell geometry as close as possible to the geometry of the expected anomalies. This

point is the main reason for the MST. For instance, in many real applications the depth

variation of the velocity may be significantly greater than the lateral variation. Hence, it

makes more sense to have the 2-D MST model partition for each layer, rather than for the

whole 3-D model. This is the case for the Southern California example shown later. Also in

the 2-D reflection tomography example shown later, because that we try to determine the

geometry of each model interface and there is no connection between any two interfaces, it

would be reasonable to have MST model partition for each layer rather than the whole

model.

剩余223页未读，继续阅读

minase

粉丝: 1
资源: 3

地震层析成像原理与方法

bh_tomo层析成像

地球物理matlab源代码-PFAST:传输和反射地震旅行时间层析成像的高性能软件包

tomosar_ml:层析成像SAR数据的机器学习

TOMO2GMT.package.tar.gz_mgt_地震层析_地震层析成像_层析 地震_层析成像

ray2mesh-1.0.tar.gz_地震_地震 成像_地震层析成像_地震成像_层析成像

fmtomov1.0.tar.gz_fmtomov1.0.tar_三维层析成像_地震层析成像_层析 地震_层析三维成像

FATOMO.package.tar.rar_fatomo_seismic_地震_地震层析_地震层析成像

地震层析成像 可以改动代码

MCTomo:2D和3D蒙特卡洛地震层析成像

鸟取MJMA 7.3级地震震源区地震层析成像研究 (2005年)

最新资源

TOMO2GMT.package.tar.gz_mgt_地震层析_地震层析成像_层析地震_层析成像

ray2mesh-1.0.tar.gz_地震_地震成像_地震层析成像_地震成像_层析成像

fmtomov1.0.tar.gz_fmtomov1.0.tar_三维层析成像_地震层析成像_层析地震_层析三维成像

地震层析成像可以改动代码