周敏,武杰.不依赖于子波的声波多尺度全波形反演方法[J].声学技术,2018,37(6):521~527 |
不依赖于子波的声波多尺度全波形反演方法 |
Acoustic source-independent multi-scale full waveform inversion |
投稿时间:2018-05-22 修订日期:2018-06-30 |
DOI:10.16300/j.cnki.1000-3630.2018.06.002 |
中文关键词: 速度场建模|全波形反演|不依赖于子波|多尺度反演 |
英文关键词: velocity model building|full waveform inversion|source-independent|multi-scale inversion strategy |
基金项目:国家自然科学基金资助项目(41574106)、国家科技重大专项(2017ZX05008-008-041)、国家重大科研装备研制项目(ZDYZ2012-1-05-03)。 |
|
摘要点击次数: 960 |
全文下载次数: 724 |
中文摘要: |
伴随着油气勘探目标复杂性的日益增加,对速度场建模和成像方法都提出了更高的要求。声波介质下的全波形反演方法是现阶段精度最高的速度场建模方法,在复杂介质速度场建模方面具有一定的应用前景。然而,传统的波形反演方法对地震子波的准确性具有较高的依赖性,并且在地震数据主频较高的情况下难以得到有效的反演结果。为此,基于声波方程,通过修改维纳滤波器,通过利用参考道构建滤波器,提出了一种不依赖于子波的全波形反演方法,能够有效避免波形反演对子波的依赖性。由于滤波器目标子波选取较为自由,该方法可以与多尺度反演策略有机结合,实现频率由低到高的递进反演,能够进一步提高反演的稳定性。理论分析和模型试算证明,采用不依赖于子波的多尺度反演方法能够有效避免子波提取问题,在子波错误的情况下可以得到准确的反演结果,并且能够在保证反演稳定性的基础上提高反演精度,反演效果优于传统的波形反演方法。 |
英文摘要: |
As the most accurate and promising model building method, acoustic full waveform inversion is expected to provide a more effective velocity model. However, full waveform inversion is a data-driving inversion method, which needs accurate source wavelet term and low frequency data. In this paper, a source independent inversion strategy is proposed by modifying the Wiener filter using the reference trace. By filtering the modeled and observed data, the source term can be eliminated from the misfit function. Besides, because the target wavelet in the modified Wiener filter can be chosen relatively freely, the new method is naturally combined with the frequency based multi-scale inversion strategy to further improve the stability of full waveform inversion. The proposed method is applied in practical application to verify its correctness and efficiency. The theoretical analysis and numerical test illustrate that the new method can avoid the source estimating procedure and improve the inversion accuracy and stability of full waveform inversion, which outperforms the conventional full waveform inversion strategy. |
查看全文
查看/发表评论 下载PDF阅读器 |
关闭 |