Data processing: artificial intelligence – Fuzzy logic hardware – Fuzzy neural network
Reexamination Certificate
1999-11-30
2003-09-30
Follansbee, John (Department: 2121)
Data processing: artificial intelligence
Fuzzy logic hardware
Fuzzy neural network
C706S016000, C706S929000
Reexamination Certificate
active
06629086
ABSTRACT:
BACKGROUND OF THE INVENTION
Field of the Invention
The present invention relates to interpretation of geochemical measurements obtained by pyrolysis of rock samples.
Description of the Prior Art
One of the objectives of organic geochemistry in oil exploration consists in characterizing the organic matter contained in source rocks. The Rock-Eval pyrolysis method (trademark registered by Institut Francais du Pétrole) has been designed to answer this need. This method allows obtaining a series of measurements which allow satisfactorily evaluation of the petroleum potential of the source rock, the amounts of free hydrocarbons contained therein, as well as the type and the maturation state of the organic matter.
Rock-Eval pyrolysis is a fast and inexpensive method allowing access to the characteristics of the organic matter. The Rock-Eval method consists in pyrolizing rock samples by heating them according to a well-determined temperature program. Pyrolysis of rock samples provides a series of parameters that are used for characterizing the organic matter contained in the pyrolysed samples.
U.S. Pat. Nos. 4,153,415, 4,352,673 and 4,519,983 illustrate the Rock-Eval method. Chapter 11.2 “Screening Techniques for Source Rock Evaluation” concerning “Rock-Eval Pyrolysis” by J. Espitalié and M. L. Bordenave in “Applied Petroleum Geochemistry”, 1993, Editions Technip, France, also describes the Rock-Eval method.
SUMMARY OF THE INVENTION
The present method comprises stages using computational intelligence (IC) techniques for automatic analysis of the measurements obtained, notably through the Rock-Eval pyrolysis method. The method allows fast and reliable description of the main characteristics of the organic matter contained in source rocks.
The method is based on integration of two computational intelligence techniques, i.e. artificial neural networks and fuzzy sets. The suitability of these techniques for interpretation of the data obtained from Rock-Eval pyrolysis has allowed obtaining rapidly diagnoses that are close to and, in some cases, more accurate than diagnoses that would be made by a human expert.
The user can also have access to type samples close to those being studied.
In a variant, the analysis method allows global study of the evolution of the organic matter all along wellbores. During this analysis, the system calculates correlated relations between various geochemical data. These correlations allow the user to follow the change in the organic matter during the maturation process. The case of wellbores containing highly evolved organic matter can also be studied in order to evaluate the initial petroleum potential and the amounts of hydrocarbons that could migrate during the evolution.
The present invention thus relates to a method intended for automatic interpretation of geochemical measurements obtained by pyrolysis of a rock sample in order to obtain characterization of the organic matter contained in the sample. According to the invention, the following steps are carried out:
using rock samples having known petroleum characteristics in order to train an artificial neural network;
using the neural network to obtain parameters which pertain to the organic matter of a rock sample; and using fuzzy sets for refining interpretation of the parameters at a network output of the neural network.
The organic matter of the sample can be characterized by determining at least the type and the maturation state of the organic matter, and the petroleum potential.
An evolved series of rock samples taken during a single drilling operation can be analyzed by carrying out the following complementary stages:
from knowledge of the type of the organic matter contained in the samples, determining the correlation function f connecting the hydrogen index (HI) to the maximum pyrolysis temperature (Tmax) for the reference series of the organic matter type, determining the correlation function g connecting the hydrogen index of the evolved series to the hydrogen index of the reference series,
using g and f to obtain the values of the hydrogen index as a function of depth,
estimating the initial petroleum potential.
The correlation functions can be established by means of multilayer neural networks.
REFERENCES:
patent: 5751915 (1998-05-01), Werbos
Shin-ichi Horikawa et al; On Fuzzy Modeling Using Fuzzy Neural Networks with the Back-Propagation Algorithm; 1992; IEEE; 9201454; 801-806.*
Masoud Nikravesh et al: Dividing Oil Fields into regions with Similar Characteristic Behavior Using Neural Network and Fuzzy Logic Approaches; 1996; IEEE; 0-7803-3225-3-6/96; 164-169.
Braunschweig Bertrand
Lafargue Eric
Naanaa Wady
Antonelli Terry Stout & Kraus LLP
Follansbee John
Hirl Joseph P.
Institut Francais du Pe'trole
LandOfFree
Method for interpreting petroleum characteristics of... 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 interpreting petroleum characteristics of..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method for interpreting petroleum characteristics of... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3100526