Back to Search Start Over

Reverse time migration imaging of tunnels via the finite element method using an unstructured mesh

Authors :
Fei Cheng
Yi-Fan Huang
Huai-Jie Yang
Jing Wang
Jiangping Liu
Source :
Applied Geophysics. 17:267-276
Publication Year :
2020
Publisher :
Springer Science and Business Media LLC, 2020.

Abstract

Wavefield extrapolation is critical in reverse time migration (RTM). The finite difference method is primarily used to achieve wavefield extrapolation in case of the RTM imaging of tunnels. However, complex tunnel models, including those for karsts and fault fracture zones, are constructed using regular grids with straight curves, which can cause numerical dispersion and reduce the imaging accuracy. In this study, wavefield extrapolation was conducted for tunnel RTM using the finite element method, wherein an unstructured mesh was considered to be the body-fitted partition in a complex model. Further, a Poynting vector calculation equation suitable for the unstructured mesh considered in the finite element method was established to suppress the interference owing to low-frequency noise. The tunnel space was considered during wavefield extrapolation to suppress the mirror artifacts based on the flexibility of mesh generation. Finally, the influence of the survey layouts (one and two sidewalls) on the tunnel imaging results was investigated. The RTM results obtained for a simple tunnel model with an inclined interface demonstrate that the method based on unstructured meshes can effectively suppress the low-frequency noise and mirror artifacts, obtaining clear imaging results. Furthermore, the two-sidewall tunnel survey layout can be used to accurately obtain the real position of the inclined interface ahead of the tunnel face. The complex tunnel numerical modeling and actual data migration results denote the effectiveness of the finite element method in which an unstructured mesh is used.

Details

ISSN :
19930658 and 16727975
Volume :
17
Database :
OpenAIRE
Journal :
Applied Geophysics
Accession number :
edsair.doi...........0c543d9e4179a9e80fd0acaf733d389c