SEPARATION OF THE DIFFRACTED AND REFLECTED WAVEFIELDS BY

SEPARATION OF THE DIFFRACTED AND REFLECTED WAVEFIELDS BY F-K FILTERING Brydon Lowney 1*, Ivan Lokmer 1, Chris Bean 2, Gareth O’Brien 3, Michael Igoe 3 1 Irish Centre for Research in Applied Geosciences, University College Dublin, Stillorgan Rd, Belfield, Dublin 4, Ireland, 2 School of Geophysics, Dublin Institute for Advanced Studies, 5 Merrion Square, Dublin 2, Ireland, 3 Tullow Oil Ltd, Central Park, 1, Carmanhall and Leopardstown, Dublin 18. We would like to thank Shearwater for providing a license to Open. CPS software which was used in this work Corresponding author email: brydon. lowney@icrag-centre. org Abstract Diffractions of seismic waves can be a powerful tool in a geophysicist’s arsenal. Diffraction images can enhance features the standard seismic resolution, helping identify features such as faults, pinch-outs, and fractures. However, in the conventional processing workflow diffractions are not utilised. This paper aims to demonstrate diffraction imaging through a novel technique which suppresses reflections through filtering in the f-k domain. Herein, the f-k method is proposed and then applied to a synthetic and real dataset. In the f-k domain, events are separated based upon their geometry and their velocity. As such, reflection events (which appear linear in the f-k domain) can be filtered, leaving the diffractions (which form fan shapes). These can then be migrated using a standard migration scheme to create a diffraction image. This technique was applied to both a synthetic and a real dataset alongside a plane wave destructor technique for comparison. It effectively suppressed the reflections from the wavefield, allowing the diffractions to be imaged. This was used in conjunction with a conventional reflection image to highlight channel discontinuities, faults, and breaks in geological horizons in the real dataset. This project is funded by the Petroleum Infrastructure Programme (PIP). Theory 1 To begin the process of f-k filtering the data is first stacked to create a conventional poststack image. This image is then converted to the f-k domain through a 2 D Fourier Transform (2 DFT). In this domain, the reflectors are separated by their apparent velocity (Figure 1) forming linear shapes, while diffractors form fans of energy. 3 2 Methodology As the reflections and diffractions are separated in the f-k domain, a band-stop filter can be designed to suppress the reflections whilst retaining as much diffraction energy as possible (Figure 2). Figure 1: Time section (left) and the corresponding f-k domain (right) showing how reflections (V 1 -4) and diffractions (V 4) appear in f-k. Synthetic Tests With the reflections suppressed, the data can be migrated to create a diffraction image (Figure 3). The original post-stack image can also be migrated to create a conventional image which is used for comparison. The two images can then be combined to create an enhanced resolution image. Figure 2: Synthetic data tests of the f-k filter showing the unmigrated stack (top left) and the corresponding f-k (top right), as well as the post-filtered image (bottom left) and the corresponding f-k (bottom right). 4 Figure 3: Migrated conventional image (left), f-k filtered image (middle), and combined image (right). Note how the actual location of the pinch-out point (red circle) can be identified by a bright spot of energy, as opposed to the apparent pinch-out point (green) seen from just the conventional image, these points are approximately 75 m apart. This technique was applied to the real data (supplied by Tullow Oil) to show significant enhancements in horizon discontinuities, faults, and channels (Figure 4). This was compared with the existing plane-wave destructor (PWD) technique and some zoomed images have also been provided to show the advantages of the f-k filter. Future Work Here we have demonstrated an alternate method of diffraction imaging which improves the seismic image. Future work will focus on refinement of the f-k filter method and applying the method to identify pinch-outs and thief sands, on top of applying the technique to data from the Irish offshore. This publication has emanated from research supported in part by a research grant from Science Foundation Ireland (SFI) under Grant Number 13/RC/2092 and co-funded under the European Regional Development Fund and by PIPCO RSG and its member companies. Conventional f-k Combined PWD Combined 4 b PWD Combined f-k Filtered Conventional 4 a 5 Real Data Tests Figure 4 a: Conventional seismic image (top), f-k filtered diffraction image (middle), and combined image (bottom). Note, the clear improvements on the horizons (showing more of the discontinuities), the channels, and the faults underlying the channels. 2 areas have been highlighted in particular in 4 b, shown by the red and green boxes. Figure 4 b: Conventional image (top), f-k filtered combination image (middle), and PWD filtered combination image. The f-k filtered image clearly displays the channel shape on the left, as well as the faults underlying the channels on the right image when compared with PWD.