Data processing: measuring – calibrating – or testing – Measurement system in a specific environment – Earth science
Patent
1997-08-04
1999-08-17
McElheny, Jr., Donald E.
Data processing: measuring, calibrating, or testing
Measurement system in a specific environment
Earth science
706929, G06F 1900
Patent
active
059407778
DESCRIPTION:
BRIEF SUMMARY
BACKGROUND OF THE INVENTION
The present invention relates to a method for automatic recognition of seismic facies between two horizons, or about a horizon of a geological area, or, more particularly between two horizons or about a horizon defined on a seismic section associated with said geological area.
At present, nearly all geological and geophysical interpretations relative to seismic facies are carried out on an interpretation station and belong to the specialized domain of seismic stratigraphy.
In seismic stratigraphy, it is customary to identify and to represent on a map the variations of seismic facies in a given slice of the geological area to be surveyed (called mapping). The slice may or may not be between two marked horizons.
A seismic facies unit is a group of seismic reflections having configurations, i.e., external shape and internal parameters, which are different from one facies unit to another. The configurations may also be different between two adjacent or consecutive facies units.
The seismic facies units are usually defined by analyzing three families of parameters: etc.), etc.).
The recognition of the seismic facies in a given geological area is very important because it provides useful information, particularly about the types of sedimentary deposits and the anticipated lithology.
To succeed in recognizing the seismic facies of a given geological area, it is necessary to define each of them first by separately analyzing at least each of the above-mentioned three families of parameters. Next, the parameters should be synthesized in order to gather the maximum data or information about the seismic facies present in the geological area.
The cost of such an analysis and the means to be employed, particularly the data processing means, are excessively high as compared to the results obtained.
In fact, if the seismic facies which one wishes to recognize belong to stratigraphic pinchouts and/or to turbiditic channels, it is very difficult to discriminate between the anomalies when they appear on the usual seismic sections, even if those anomalies are recognized by the well seismic survey as being present in the area concerned, provided that a well is available in the area, which may not be the case.
In EP-0 561 492, a method is described for improving the well logging by making use of neural networks. The particular network described is a layered network. From a statistical standpoint, a layered network is a universal approximator of the boundaries between classes, but, above all it is, a supervised network. In other words, the quantity obtained in the output of the neural network is compared with a quantity known and determined by other methods, until a coincidence or quasi-coincidence is obtained between the quantities.
Since the topological maps due to Kohonen are used in other fields, particularly in the medical field to determine models susceptible to imitate a number of the functions of the brain by reproducing some of its basic structures, geophysicists have attempted to apply them to the field of geophysics.
Particular applications are described in U.S. Pat. No. 5,373,486, which deals with the classification of seismic events by using Kohonen antagonistic networks, in U.S. Pat. No. 5,355,313, which describes the interpretation of aeromagnetic data, and in U.S. Pat. No. 5,181,171, which describes an interactive neural network adapted to detect the first arrivals on the seismic traces.
SUMMARY OF THE INVENTION
It is an object of the present invention to propose a method for recognizing seismic facies from a seismic section associated with a geological area, and to do this automatically via an unsupervised neural network.
The present invention relates to a method for recognizing seismic facies between two horizons or about a horizon of a geological area, and comprises the steps of: and determining the value of said parameter for each of the trace portions of the set, facies to be recognized, each cell being assigned a value of the recognition parameter, portions, so that, when the
REFERENCES:
patent: 5181171 (1993-01-01), McCormack
patent: 5355313 (1994-10-01), Moll et al.
patent: 5373486 (1994-12-01), Dowla et al.
J. Parikh, et al., "Selective detection of linear features in geological remote sensing data", Proceedings of SPIE: Applications of Artificial Neural Networks III, vol. 1709, No. 2, Apr. 21-24, 1992, Orlando, Florida, pp. 963-972.
W. Maurer, et al., "Seismic event interpretation using self-organizing neural networks", Proc. SPIE: Applications of Artificial Neural Networks III, vol. 1709, No. 2, Apr. 21-24, 1992, Orlando, Florida, pp. 950-958.
K. Y. Huang, et al., "A hybrid neural network for seismic pattern recognition", IJCNn International Joint Conference on Neural Networks, vol. 3, Jun. 7-11, 1992, Baltimore, Maryland, pp. 736-741.
Elf Aquitaine Production
McElheny Jr. Donald E.
LandOfFree
Automatic seismic pattern recognition method does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Automatic seismic pattern recognition method, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Automatic seismic pattern recognition method will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-325460