Data processing: measuring – calibrating – or testing – Measurement system in a specific environment – Earth science
Reexamination Certificate
2002-03-04
2003-06-10
Lefkowitz, Edward (Department: 2862)
Data processing: measuring, calibrating, or testing
Measurement system in a specific environment
Earth science
C703S005000
Reexamination Certificate
active
06577955
ABSTRACT:
PRIORITY
This application claims the benefit of French application number 0102962, filed on Mar. 5, 2001, and is an English translation thereof.
FIELD OF INVENTION
The present invention relates to methods of tomographically inverting picked events on seismic traces.
BACKGROUND OF THE INVENTION
To reconstruct images of the subsoil, geologists or geophysicists conventionally use sound emitters placed on the surface, for example. Such emitters emit waves which propagate through the subsoil and which are reflected on the surfaces of the various layers thereof (reflectors). Sound waves reflected to the surface are recorded as a function of time by receivers. The signals received by the receivers are known as seismic traces.
It is conventional to pick portions of such seismic traces which correspond to reflections of sound pulses emitted from the surface, and which correspond to reflectors of interest, and also to determine the travel times that correspond to such reflections.
Tomographic inversion techniques consist in modeling velocity fields within the subsoil as a function of the acquired seismic traces and of selected events.
Numerous inversion techniques are already known.
It is conventional to invert reflected wave arrival times, and in particular to determine the velocity field in the subsoil by minimizing the differences between the observed arrival times and the arrival times predicted b, the models, which times are calculated in the inversion model by tracing rays between the emitters and the sources
Travel time inversion is difficult to implement association with geographical structures that are complex, in particular because of the lack of discrimination between multiple arrivals.
Other tomographic inversion techniques are techniques which make use of depth migration.
Depth migration prior to adding seismic data consists in determining, for each surface position (x,y), a collection of migrated traces conveying events which describe the subsoil vertically below such surface position (x,y). Such a collection of traces is generally ordered in classes of source-receiver distances (distances also referred to as “offset” distances), and also by classes of increasing specular angles, and more rarely by classes of source-receiver segment orientation.
When the velocity model used is a proper reflection of reality, then the depth associated with an event that is reflected vertically below a surface position is substantially constant regardless of the offset distance (or indeed the specular angle) of the trace in question.
Methods of tomographic inversion in depth use this characteristic to define the inversion criterion that is to be minimized.
In particular, an inversion method has been proposed in “Velocity analysis by iterative profile migration” by Kamal Al-Yahaya, Geophysics, Vol. 54, No. 6, 1989, pp. 718-729 in which the velocity model of the subsoil is determined by minimizing the departure from horizontal of the depth/distance curve for a collection of migrated traces.
To be able to implement that technique, it is necessary for a plurality of depth migrations to be performed, which is particularly troublesome to implement when several iterations are needed for convergence.
SUMMARY OF THE INVENTION
An object of the invention is to propose an inversion technique which significantly reduces the number of depth migrations to be performed in order to construct an optimum model, and which does not require large amounts of computer power.
To this end, the invention proposes a method of updating a subsoil velocity model, in which depth migration prior to data addition is implemented on a set of seismic traces acquired in register with said subsoil, the migration being implemented with the help of at least one starting velocity model and serving to determine one or more trace collections each describing the subsoil vertically below a point on the surface; one or more events which reflect vertically below the surface point under consideration are picked on at least one collection of traces obtained using said migration, and for each picked event a reflector depth and dip are determined, as is a reflector dip in register with said surface point; and ray tracing is implemented between said reflector and the surface to determine collections of sound source and receiver pairs, together with data characteristic of travel times and time gradients which correspond to the ray traces associated with said pairs; wherein subsequent processing is implemented in which the following steps are iterated: dynamically migrating travel time and time gradient data previously obtained with the help of paramet
6
rization of the velocity field; characterizing the alignment of the migrated points obtained in this way; and updating the parameterization; the processing selecting the velocity field parameterization which optimizes the alignment of said migrated points.
The above three steps can be repeated until an alignment that is judged to be satisfactory is obtained.
It should be observed that the processing proposed in this way for selecting the parameterization of the velocity field does not require seismic data to be migrated in depth (which is very expensive) every time after each velocity update.
Consequently, it will be understood that the method proposed by the invention can be implemented without requiring large amounts of computer power.
In an advantageous implementation, the depth migration prior to addition of the seismic data is implemented for a plurality of velocity fields (common reflection point [CRP] scan, for example), and in order to determine the reflector depth and dip, that one of the velocity fields is selected which minimizes the departure from the horizontal of the depth/offset distance curve or of the depth/specular angle curve.
Furthermore, the depth of the reflector can advantageously be determined for different offsets or specular angles.
In one possible preferred implementation, in order to select a velocity field parameterization which optimizes the alignment of the points that are migrated from the time and time gradient seismic data, the distance is determined between the reflector portions seen for different offsets of the collection, with the selected velocity field parameterization being that which minimizes said distance.
In another implementation, in order to select a velocity field parameterization which optimizes the alignment of the points which are migrated from the time and time gradient data, the depth differences between the intercept points between a given vertical and previously updated reflector portions parallel to the reflector and which pass through the new migrated points, with the selected velocity field parameterization being that which minimizes said difference.
In another variant, the dips considered for each offset of a given collection need not be parallel, in which case they can be determined from the rays.
It is also possible to implement the following steps in order to select a parameterization for the velocity field: a local reflector passing through a migrated point is determined; the other rays of the collection are traced between said local reflector and the surface; a time error is determined which is a function of the difference between the travel time corresponding to at least one of said rays and the travel time of the trace which, in the collection, corresponds to the same offset distance or the same specular angle; and a velocity field parameterization is selected which minimizes said time error and optimizes the alignment of the points which are migrated from the time and time gradient seismic data.
In a variant, or in addition, it is also possible to implement the following steps: a reflector is determined passing through a migrated point; the other rays of the collection are traced between said reflector and the surface; a time error is determined which is a function of the difference between the travel time gradient corresponding to at least one of said rays and the travel time gradient of the trace which, in the collec
Compagnie Generale de Geophysique
Duane Morris LLP
Lefkowitz Edward
Taylor Victor J.
LandOfFree
Methods of tomographically inverting picked events on... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Methods of tomographically inverting picked events on..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Methods of tomographically inverting picked events on... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3141552