...
首页> 外文期刊>物理探查 >Numerical experiments for Full-wave tomogrpahy (2)-Source inversion
【24h】

Numerical experiments for Full-wave tomogrpahy (2)-Source inversion

机译:全波断层扫描(2) - 源反转的数值实验

获取原文
获取原文并翻译 | 示例

摘要

Seismic waveform is influenced not only by medium property but by source parameters, hence source parameter inference is important task in Full-wave tomography (FWT). As discussed in Tsuchiya (1997a), the mechanism of explosive borehole source can be described by the diagonal components of moment density tensor; two time series of vertical and in-plane horizontal vector dipoles, in case of 2D FWT. Moreover it can he assumed that they are determined by the combination of the normalized common time series (source function) and two maximum values of each vector dipole (MVD's).While we have some possibility for prior estimate of the source function, the direct measurement of directivity parameter (the ratio of two MVD's) is difficult. This fact suggests us to infer the source parameters through the FWT calculation.Because the source inversion as well as the property parameter inversion in FWT is a nonlinear problem, the method proposed by Tsuchiya (1997a) in which the source parameters are infered jointly with the property parameters is naturally thought out. Thus the numerical experiments to study practical procedure of his method were implemented and are discussed in this paper.Conclusions are as follows. (1) Accurate evaluation of source parameters is important for inference of Vp and Vs distribution. (2) Joint inversion of source parameters with Vp and Vs is effective. (3) Source parameter inversion in terms of the gradient method tends to converge towards local minimum. This effect is mainly caused by coupling of the prior errors of MVD's and source function. (4) One possible way to solve this problem is to deal with the averaged time series as unknown until the property parameters are sufficiently restored.
机译:地震波形不仅受媒体属性而且源参数影响,因此源参数推断是全波断层扫描(FWT)中的重要任务。如Tsuchiya(1997A)中所讨论的,爆炸性钻孔源的机理可以通过力矩密度张量的对角线组分来描述;两个时间序列的垂直和面内水平矢量偶极子,在2D FWT的情况下。此外,他可以假设它们是由归一化的公共时间序列(源函数)的组合来确定每个矢量偶极子(MVD)的两个最大值。虽然先前估计源功能的可能性,直接测量方向性参数(两个MVD的比例)很难。这一事实建议我们通过FWT计算推断源参数。因为源反转以及FWT中的属性参数反转是非线性问题,所以TSUCHIYA(1997A)提出的方法,其中源参数与源参数共同推断出来物业参数自然认为。因此,实施了研究其方法实际过程的数值实验,并在本文中讨论。结论如下。 (1)对源参数的准确评估对于VP和VS分布的推动是重要的。 (2)具有VP和VS的源参数的联合反演是有效的。 (3)源参数反转在梯度法方面倾向于达到局部最小值。这种效果主要是由MVD和源功能的先前误差耦合引起的。 (4)解决此问题的一种可能方法是将平均时间序列处理为未知,直到属性参数充分恢复。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号