Developing and Analyzing Green's Theorem Methods to Satisfy Prerequisites of Inverse Scattering Series Multiple Attenuation for Different Types of Marine Acquisition: Towards Extending Prerequisite Satisfaction Methods for On-Shore Exploration
Date
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Inverse Scattering Series (ISS) algorithm can directly achieve the objectives of seismic processing without requiring any subsurface information. For achieving the potential capabilities of ISS algorithm, there are prerequisites that need to be satisfied. These prerequisites (including separating the reference wave from the reflected data, estimation of the source wavelet, and deghosting) can be satisfied by using Green's-theorem methods. This dissertation provides three contributions in satisfying the prerequisites for Inverse Scattering Series (ISS) multiple removal algorithm.
Chapter 2 examines the impact of a specific seismic-acquisition design (over/under cables) on the wave-separation methods. When the depth difference between the two cables is smaller, the wave-separation results are more accurate and have less errors. In the (x, ω) domain, Green's theorem requires the prediction point to be chosen away from the measurement cable, but it can accommodate a non-at cable (e.g., at ocean bottom). Green's theorem in the (k, ω) domain can predict the separated wavefields on the cable. However, it requires a flat cable to perform Fourier transform over the measurement surface.
Chapter 3 presents a method for determining the correct reference velocities. The criteria for finding the correct reference velocities could be the invariances of source wavelet at different output points below the cable for the point source data, or the invariances along one radiation angle for the source array data.
The third project investigates and compares three different wavelet estimation methods, including: (1) the Wiener filter method, (2) the spectral division method, and (3) the Green's-theorem method. Comparing with the other two methods, the Green's-theorem method demonstrates strength when the data contains random noise, since it utilizes an integral along the measurement surface, which tends to reduce random noise.