Detection of Fracture Networks and Faults Based on Multiply Scattered Waves

Fractures and faults are widely seen in nature. They can play important roles as storage for natural resources, paths for the migration of fluids including hydrocarbons, structural traps for petroleum reservoirs. Fracture characterization and fault imaging by using seismic methods have contributed greatly in finding economic reservoirs. In this dissertation, we focus on how to use multiply scattered waves to characterize fractures and image the faults.

The first part of the dissertation focuses on the fracture characterization using the Gaussian wave packet (GWP). Conventionally, using seismic anisotropy to study fractured media has gained great success in the cases where the effective medium theory (EMT) holds. As a supplement to EMT, our proposed method characterizes the fractures using multiply scattered waves when the EMT is no longer valid.

In our proposed method, we use a GWP which approximates as a local plane wave to interact locally with the fractured medium. The resultant multiply scattered waves can be used to characterize fractures in terms of spacing, compliance and orientation. We first study the propagation of the GWP to investigate how well it behaves as a local plane wave relative to the propagating distance and heterogenous velocities. Then, we describe how to build the GWP wavefield from the wavefields of point sources. Finally, we develop the method to characterize fractures. Using numerical examples, we show the potential usefulness of our method in fracture characterization.

The second part of the dissertation is concentrated on imaging high-angle faults using multiply scattered waves. When the faults have high angles, the directly reflected waves from fault planes need a longer offset to be recorded. However, a longer offset may not be favorable in practice.

Traditionally, the high-angle faults are interpreted and/or extracted by attributes but not imaged directly. With a limited offset, we aim at using multiply scattered waves to image the high-angle faults directly.

To fulfill the goal of imaging faults directly, we develop the asymmetrical reverse time migration (asym-RTM) algorithm. The asym-RTM images the sub-horizontal reflectors in its first iteration. Then, the imaged reflectors are added into the velocity model for the second iteration. In this way, we can utilize the second-order scattered waves to image the faults directly.