Enhancing internal multiple prediction by using the inverse-scattering series: Methodology and field application

Abstract

We introduce four approaches that dramatically enhance the application of the inverse scattering series method for field data internal multiple prediction. The first approach aims to tackle challenges related to input data conditioning and interpolation. We addressed this through an efficient and fit-for-purpose data regularization strategy, which in this work was a nearest-neighbor search followed by differential moveout to accommodate various acquisition configurations. The second approach addresses cost challenges through applying angle constraints over both the dip angle and opening angle, reducing computational cost without compromising the model’s quality. We also propose an automatic solution for parameterization. The third approach segments the prediction by limiting the range of the multiple’s generator, which can benefit the subsequent adaptive subtraction. The fourth approach works on improving predicted model quality. The strategy includes correctly incorporating the 3D source effect and obliquity factor to enhance the amplitude fidelity of the predicted multiples in terms of frequency spectrum and angle information. We illustrate challenges and report on the improvements in cost, quality or both from the new innovative approaches, using examples from synthetic data and from three field data 2D lines representative of shallow and of deep water environments.

Application of deconvolution interferometry to receiver ghost reflection in vertical cable seismic survey

“…Therefore, it is an important work to suppress multiples in seismic data processing, especially for marine seismic exploration. Various methods have been developed for multiples attenuation, such as surface-related multiple elimination (Verschuur et al,1992;Dragoset et al, 2010), inverse scattering series (Weglein et al, 1997;Wu et al, 2021), Radon transform (Foster & Mosher, 1992;Nowak & Imhof, 2006), multiple prediction through inversion (Wang, 2007), Marchenko method (Broggini et al, 2012;Wapenaar et al, 2014;Staring & Wapenaar, 2020) and so on.…”