Data processing: measuring – calibrating – or testing – Measurement system in a specific environment
Reexamination Certificate
1999-10-18
2003-01-14
Lefkowitz, Edward (Department: 2862)
Data processing: measuring, calibrating, or testing
Measurement system in a specific environment
C702S014000
Reexamination Certificate
active
06507787
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates to a method for the suppression of multiple reflections from marine seismic data, more specifically, to a method for the near offset range suppression of multiple reflections where the simulation of multiple reflections by means of a Kirchhoff-type algorithm is applied to a common near-offset section. As a result, significant improvement in the seismic image is achieved.
BACKGROUND INFORMATION
Routine interpretation of processed seismic data is mainly focused on understanding primary reflections. Multiples, however, also occur in all seismic reflection data, acquired both in land and marine environments. Although multiples may bring useful data on the subsurface medium, present processing technology is still not able to make good use of this information. On the contrary, multiple arrivals are seen as undesirable events.
If not properly attenuated or removed, multiples can interfere with primary reflections, leading to difficulties in the estimation of velocities and amplitudes. In many land data cases, multiple attenuation can be adequately achieved by deconvolution and CMP stacking.
Because of the presence of strong reflections on the water-air contact (free surface), in particular sea-bottom multiple reflections, marine data most often require additional processing.
In general, methods of multiple attenuation or suppression can be divided into two categories, namely (a) filtering—based on some characteristics of the multiple energy involved and (b) prediction—simulation of multiples and subtraction from the original data. The well known Radon and f-k transform algorithms, generally applied in the CMP domain, belong to the first group. An internal mute on near offsets has to be applied, as a discrimination between primaries and multiples is more difficult in this range. Algorithms of the second category are based on wave-propagation approaches such as forward modeling or inverse scattering. These are generally expensive and require, moreover, a good estimate of the source signature for useful results.
The increasing use of seismic data in reservoir characterization, together with the large number of exploratory prospects on targets of stratigraphic nature, poses as a very desirable aim any improvement in resolution that can be achieved by seismic processing. In this context, accurate elimination of multiple reflections in near-offset range within a target region may be crucial to correctly image, say, a fine stratigraphic sequence or to derive reliable seismic attributes of a reservoir.
In the case of deep-water data, suppression of first-order sea-bottom multiple and peg-leg reflections (i.e., which reverberate only once in the water layer) is of great importance. These rather strong multiples may have the same traveltime as the primary reflections of target reflectors.
U.S. Pat. No. 4,665,510 teaches a method of attenuating multiple reflections in a conventional seismic processing flow. Predictive Deconvolution After Stack is used to reduce the multiple interference on stack velocity analysis. The Deconvolution After Stack emphasizes the primary reflections with their stack velocity ranges, and this information can aid in selecting a good multiple stack velocity for filtering in F-K domain.
U.S. Pat. No. 5,729,506 teaches a computationally economical method for applying multi-dimensional multiple-reflection attenuation to a marine seismic-signal data set comprised of a plurality of common shot sections that have surface-multiple wavefields embedded therein. A common shot section is chosen from among the plurality of common shot sections. The predicted surface-multiple wavefield nearest to the chosen common shot section is adaptively filtered to match the predicted surface-multiple wavefield to the surface-multiple wavefield embedded in the chosen common shot section. The matched predicted surface multiple wavefield is subtracted from the embedded surface multiple wavefield to provide a multiple-free common shot section.
U.S. Pat. No. 5,051,960 teaches a method of removing multiple reflection events from seismic records. A shallow primary reflection event is used as a model for a deeper primary, and is distorted responsive to the predicted static correction to yield a model for the multiple. The method is applied to a common depth point (CDP) record, in which the common-midpoint contribution of the multiple to each CDP record can be assumed to be equal. The technique taught in said US patent is applied solely to removal of multiples which arise from reflection at the surface during land exploration.
Hubral, P. et al in “A unified approach to 3-D seismic reflection imaging. Part I: Basic Concepts”, Geophysics vol. 61 n
o
3 p.742-758 (1996) state that the basic idea of the unified approach is to cascade Kirchhoff migration and demigration integrals to solve a variety of seismic imaging problems. The application of both integrals in sequence leads to weighted summation along certain stacking surfaces. This kind of algorithm is known in the technique as a Kirchhoff-type algorithm. The present invention teaches one possible application of the principles discussed in said article.
Tygel, M. et al. in “A unified approach to 3-D seismic reflection imaging. Part II: Theory”, Geophysics vol. 61 n
o
3 p. 759-775 (1996) discuss the theoretical principles stated in Part I of the article.
Weglein, A. B, in “Multiple attenuation: an overview of recent advances and the road ahead (1999)”, The Leading Edge, January 1999, pages 40-44, presents an overview of the current state of multiple attenuation as well as recent advances in the field. Filtering as well as prediction methods such as wavefield prediction and subtraction, wavefield extrapolation, free-surface multiple elimination, feedback and inverse-scattering methods are discussed.
Filpo Ferreira da Silva, E. and Tygel, M., in “Deep-water multiple suppression in the near-offset range”—The Leading Edge, January 1999, pages 81-84, present a method to suppress first-order and peg-leg multiples in deep-water marine data. The main steps are: i) simulation of multiples by means of a weighted Kirchhoff-type summation applied to a stacked section; ii) application of an adaptive filter to adjust the simulated multiple; and iii) subtraction of the adjusted multiple from the original data. Advantageously, the technique has no limitations on short offsets and no requirement to estimate source wavelets and reflection coefficients.
Spitz, S., in “Pattern recognition, spatial predictability, and subtraction of multiple events”, The Leading Edge, January 1999, page 55 to 58, discusses the extraction of simulated or modeled multiples from the original data set, which corresponds to step iii) of the method described by Filpo Ferreira da Silva and Tygel M. in the previous paragraph. In this article Spitz shows that the subtraction methodology is mainly concerned with the shapes of the events.
Filpo Ferreira da Silva, E. and Tygel, M., in “Near-offset multiple suppression”, Society of Exploration Geophysicists (SEG), Sixty-ninth Annual Meeting, Tulsa, USA, Oct. 31-Nov. 5, 1999, describe a simulation method to effect a transformation of a given common near-offset or stacked input section (taken as an approximation of a zero-offset section and containing all primaries and multiples), into a corresponding output section that solely consists of the multiples to be suppressed.
SUMMARY OF THE INVENTION
Broadly, the main steps of the application of the method are i) simulation of multiple reflections by means of a Kirchhoff-type summation applied to a pseudo zero-offset section. As explained hereinbefore, this is an input common near-offset of stacked section that is taken as an approximation of a zero-offset section; ii) application of an adaptive filter to adjust the simulated multiple reflections and iii) subtraction of the obtained adjusted simulated multiple reflection from the original data.
In the present invention, steps ii) and iii) are carried out according to state-of-the-art principles.
Thus, the
Filpo Ferreira Da Silva Eduardo
Tygel Martin
Lefkowitz Edward
Petroleo Brasileiro S.A.--Petrobras
Taylor Victor J.
LandOfFree
Method for the suppression of multiple reflections from... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method for the suppression of multiple reflections from..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method for the suppression of multiple reflections from... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3004903