Statistical method of classifying events linked with the physical properties of a complex medium such as the subsoil

• US 6,052,651 A
• Filed: 09/21/1998
• Issued: 04/18/2000
• Est. Priority Date: 09/22/1997
• Status: Expired due to Term

• Current Assignee: Institut Francais Du Petrole
• Original Assignee: Institut Francais Du Petrole
• Inventors: Fournier, Frederique
• Primary Examiner(s): Oda, Christine K.
• Assistant Examiner(s): Taylor, Victor J.
• Application Number:
  US09/157,572
• Time in Patent Office: 575 Days
• Field of Search: 702/14, 702/16, 307/73, 367/33
• US Class Current: 702/14
• CPC Class Codes: G01V 11/00 Prospecting or detecting by methods combining techniques covered by two or more of main groups G01V1/00 - G01V9/00
• CPC Class Codes: G06F 18/40 Software arrangements specially adapted for pattern recognition, e.g. user interfaces or toolboxes therefor

Abstract

A method of determining the position in subsoil of a population of physical phenomena linked with physical properties of the subsoil. Data are collected, for example, subsoil exploration or in-situ measurements. The phenomena are marked by points in a multivariate space and defined by an attribute vector, the coordinates of each of the points being attributes associated with the phenomena, using a function expressing a multivariate probability law appropriate for the population. Modes corresponding to higher density zones of the function are selected, the number of modes giving a number of potential losses in the population of the phenomena. For each mode a set of selected point marks the most representative events of each class. A classifying function is determined from the selected points, and the probabilities of non-selected points belonging to each of the potential classes is determined by using the determined classifying functions. The position of the physical phenomena in the subsoil is determined from the determined probabilities.

Claims

1. A method of determining the position in subsoil of a population of physical phenomena linked with physical properties of the subsoil, said method comprising the steps of:

(a) collecting data about the subsoil in order to detect potential classes of physical phenomena existing in the population, the phenomena being marked by points in a multivariate space and defined by an attribute vector, the coordinates of each of the points being attributes associated with the phenomena, using a function expressing a multivariate probability law appropriate for the population;

(b) selecting modes corresponding to higher density zones of the function, the number of modes giving a number of potential classes in the population of the phenomena, where for each mode a set of selected points marks the most representative events of each class;

(c) determining a classifying function from the selected points;

(d) determining probabilities of non-selected points belonging to each of the potential classes by using the determined classifying function; and

(e) determining the position of the physical phenomena in the subsoil from the determined probabilities.

2. A method as claimed in claim 1, wherein the phenomena are of a geological or geophysical nature.

3. A method as claimed in claim 1, wherein the phenomena relate to the production of fluids from the subsoil.

4. A method as claimed in claim 1, wherein the attributes are defined by logs measured in at least one well through the subsoil.

5. A method as claimed in claim 1, wherein the attributes are seismic attributes.

7. A method as claimed in claim 1, wherein step (b) comprises selecting modes corresponding to high-density zones by projecting the function upon an axis expressing change in events depending on the function, and selecting modes corresponding to significant peaks of the function.

8. A method as claimed in claim 1, wherein step (a) comprises exploration of the subsoil.

9. A method as claimed in claim 1, wherein step (a) comprises conducting in-site measurements in the subsoil.

6. A method as claimed in 1, wherein the attributes are defined by production measurements.