Automatic borehole sonic classification method and apparatus
AssigneeSchlumberger Technology Corporation
InventorsTing Lei, Daniel Al Choboq, Josselin Kherroubi, Lin Liang, Romain Prioul, Gurami Keretchashvili, Pontus Loviken
Life of this patent
Estimating the term…
Abstract
A general-purpose workflow for automatic borehole sonic data classification to identify data into different physical categories and logging conditions, which are traditionally manually evaluated. The workflow uses machine learning techniques and physical knowledge for data classification, including pre-processing the high-dimensional high-quality dispersion modes extracted using a recently developed physical-driven ML enabled approach.
Claims21
A method of automatically classifying sonic data comprising:
- obtaining a set of sonic data from a borehole measurement of a borehole;
- using a dispersion analysis method, processing the set of sonic data from the borehole measurement;
- determining scattered dispersion points from the processing;
- extracting the scattered dispersion points as a smooth curve through a machine-learning automatic dipole interpretation approach, wherein a fast curve, a slow curve, a fast equivalent isotropic and homogeneous curve, and a slow equivalent isotropic and homogeneous curve are used in the machine-learning automatic dipole interpretation approach;
- obtaining an equivalent isotropic and homogeneous curve for the set of sonic data from the processed borehole measurement;
- inputting the smooth curve and the equivalent isotropic and homogeneous curve into a classifier;
- and
- outputting dispersion types of the sonic data from the classifier.
A method of automatically classifying data comprising:
- obtaining a set of data from a borehole measurement;
- using a dispersion analysis method, processing the set of data from the borehole measurement;
- determining scattered dispersion points from the processing of the set of data from the borehole measurement;
- extracting the scattered dispersion points as a smooth curve through a machine-learning automatic dipole interpretation approach, wherein a fast curve, a slow curve, a fast equivalent isotropic and homogeneous curve, and a slow equivalent isotropic and homogeneous curve are used in the machine-learning automatic dipole interpretation approach;
- obtaining an equivalent isotropic and homogeneous curve for the set of data from the processed borehole measurement;
- inputting the smooth curve and the equivalent isotropic and homogeneous curve into a classifier;
- performing a synthetic dispersion analysis and feeding results to the classifier as training data;
- and
- outputting dispersion types of the data from the classifier.
A non-transitory computer readable storage medium having data stored therein representing software executable by a computer, the software including instructions for classifying data per a method, the method comprising:
- obtaining a set of data from a borehole analysis of a borehole;
- using a dispersion analysis method, processing the set of data from the borehole analysis;
- determining scattered dispersion points from the processing of the set of data from the borehole analysis;
- extracting the scattered dispersion points as a smooth curve through a machine-learning automatic dipole interpretation approach, wherein a fast curve, a slow curve, a fast equivalent isotropic and homogeneous curve, and a slow equivalent isotropic and homogeneous curve are used in the machine-learning automatic dipole interpretation approach;
- obtaining an equivalent isotropic and homogeneous curve for the set of data from the borehole analysis;
- inputting the smooth curve and the equivalent isotropic and homogeneous curve into a classifier;
- performing a synthetic dispersion analysis and feeding results to the classifier;
- and
- outputting dispersion types of the data from the classifier.
Dependencies are inferred from the claim text.