Atlantic Richfield Company

Bok S Byun

E Stuart Nelan

David D Thompson

Chi Young

A method and computer system for correcting seismic signals in a seismic survey, sensed in horizontal and vertical directions at receiver locations, is disclosed. According to the disclosed method and system, the horizontal and vertical seismic signal traces are processed in prestack gathers, after normal move-out correction. For each reflection event indicated in the prestack gathers, a performance function is evaluated over a range of zero-offset directivity and directivity slope values, using the horizontal and vertical pressure wave components sensed at the receivers. The zero-offset directivity and directivity slope values that return the maximum performance function are stored in connection with the gather and the reflection event, as indicative of the zero-offset dip and offset-dependent directivity of the reflective surface corresponding to the reflected event. These zero-offset directivity and directivity slope values are then used to correct the traces, or generate a new trace, corresponding to the pressure wave signal that is in the true reflected wave ray direction, minimizing the component of the sensed pressure wave that is transverse to the longitudinal wave ray direction. Detection and analysis of pressure-to-shear wave conversion, as well as amplitude-versus-offset analysis, is then facilitated by the present invention.

Method and system for correcting for offset-dependent directivity

A method and computer system for deriving and applying normal moveout corrections to seismic survey traces is disclosed. In each common midpoint (CMP) gather, a stacking velocity function of travel time is first determined for near-offset traces. For each reflective event, a time window about a depth point is selected. Time anisotropy and velocity anisotropy values are then determined for the depth point, for example by way of a semblance analysis applied to envelopes of the traces over the full offset range. Normal moveout corrections are then determined by a weighted sum expression of travel time (zero-offset time plus moveout correction) The weighted sum adds a first travel time expression corresponding to a Dix equation expression with a second travel time expression corresponding to a skewed hyperbolic expression that includes the time and velocity anisotropy values, with the weighting dependent upon the offset of the particular trace. Near-offset traces receive relatively higher weighting of the Dix equation expression, while far-offset traces receive relatively higher weighting of the skewed hyperbolic expression. Because higher order power expressions are eliminated according to this approach, improved moveout correction accuracy is obtained over near and far offsets, with reduced sensitivity to measurement error.

Method and system for determining normal moveout parameters for long

A method and computer system for deriving and applying normal moveout phase and time shift corrections to seismic survey traces is disclosed. Normal moveout time-shift corrections are first applied to traces in a common midpoint (CMP) or common depth point (CDP) gather, based upon stacking velocity analysis applied to envelopes of the traces. One offset in the survey, preferably the maximum offset, is selected as the reference point; the NMO time shift at this reference offset is then selected as a reference NMO time-shift. The traces in the gather are each phase-shifted based upon iterated reference phase dispersion values at the reference offset, times the ratio of the normal moveout time-shifts for the trace in the gather to that of the reference offset (raised to a power). Semblance analysis is performed to identify, for each point in normal incidence time and thus for each reflective event, the optimum reference phase dispersion value. Phase corrections, based upon the optimum reference phase dispersion value as a function of normal incidence time multiplied by the ratio of the normal moveout trace and reference time shifts, are then applied to the traces in the gather. Interpolation may be used to define the optimum reference phase dispersion for CMP gathers lying between the midpoints of analyzed gathers.

Method and system for applying dispersive normal moveout corrections to

A computer system and method for correcting seismic data for phase distortion due to the effects of the near-surface layer are disclosed. Phase error corrections are performed on multiple frequency bands, beginning with a primary band and then proceeding with adjacent frequency bands. The primary band phase constant and time lag corrections are derived in a manner in which the phase correction is applied to the correlation function prior to selection of the peak time lag value, improving the accuracy of the time lag decomposition. Corrections for the secondary bands are made in such a way that the phase corrections at a tie frequency within the overlap frequencies between each secondary band and its adjacent bands are forced to be equal. At each step, surface-consistent decomposition of the phase static components and the time lag static components is performed using a weighted least-squares approach. An alternative embodiment performs the method by determining total corrections in each band to minimize the data storage requirements.

Method and system for surface-consistent phase and time lag correction

A computer-implemented method and system of correcting seismic survey data for the effects of NMO stretch is disclosed. The method operates upon a CDP gather of seismic survey data after normal move-out correction (NMO) has been applied. A semblance analysis is used to derive a stretch coefficient .kappa. profile over the gather, as a function of time and offset. The values of the stretch coefficient profile .kappa. are used in producing a time-varying filter that is applied, preferably by way of a time-domain multiplication, to incremental windows in time for each trace, with the products added to one another to generate a corrected trace. After all of the traces in the gather are corrected, an event-weighted filter is applied to the traces by way of a correlation operation, to remove high-frequency artifacts in the corrected traces. The corrected and filtered traces are then ready for stacking and conventional processing.

Growing community of inventors

Plano, TX, United States of America

Bok S Byun