EP0840141A2
Expired / Lapsed
EP
⏱️ 0.0 years remaining
Well logging data interpretation
Assignee: Western Atlas International Inc
•
Filing: 1997-10-28
•
Granted: —
•
Expires: 2017-10-28
Abstract
Abstract A method for producing synthetic tool responses for a well logging tool for an earth formation, the method including, in one aspect, generating wellbore logging data for a particular part of an earth formation with a wellbore logging system with a wellbore logging tool, the earth formation having at least one layer, producing an input earth model of the particular part of the earth formation based on the wellbore logging data, inputting the input earth model to a trained artificial neural network, e.g. )resident in a computer, the computer with the trained artificial neural network processing the input earth model and producing synthetic tool responses for the wellbore logging tool for one point or for a plurality of points in the earth formation.
Inventors
- 👤 Alberto G. Mezzatesta
- 👤 Michael A. Jervis
- 👤 David R. Beard
- 👤 Kurt-M. Strack
- 👤 Leonty A. Tabarovksy
Bibliographic Details
Application Number
—
Jurisdiction
EP
Data Source
google_patents
Last Synchronized
2026-10-25 22:10
CPC Classifications
G; G01; G01V; G01V3/00; G01V3/38; G01V1/00; G01V1/28; G01V1/282; G01V1/40; G01V1/44; G01V1/48; G01V1/50; E; E21; E21B; E21B2200/00; E21B2200/22; G01V2210/00; G01V2210/60; G01V2210/61; G01V2210/614
Claims Analysis
Total Claims: 20
Claim 1
Independent Claim
105 words
A method for forward modeling to produce synthetic tool
responses for a well logging tool for an earth formation having at least one
layer, the well logging tool disposed in a wellbore in the earth formation, the
method comprising producing an input earth model of the earth formation based on
the wellbore logging data, inputting the input earth model to a primary trained artificial
neural network, processing the input earth model with the primary trained
artificial neural network to produce an output model that includes a set of
synthetic tool responses for the wellbore logging tool for at least one point in
the earth formation.
Claim 2
Dependent Claim
37 words
The method of claim 1 wherein the wellbore logging data is data from
the group consisting of electrical resistivity data, acoustic data, seismic data, vertical
seismic profiling data, magnetic data, gravimetric data, temperature data, and nuclear
data.
Claim 3
Dependent Claim
20 words
The method of claim 1 wherein the at least one point in the earth
formation is a plurality of points.
Claim 4
Dependent Claim
35 words
The method of claim 1 wherein the at least one layer of the earth
formation is a plurality of layers and the at least one point in the earth formation is a
plurality of points.
Claim 5
Dependent Claim
132 words
The method of claim 1 wherein the primary trained artificial neural
network is produced by a method comprising training a first artificial neural network with an input training
set, the input training set comprising a plurality of earth models based on
wellbore logging data and further comprising a set of wellbore logging tool
responses associated with each earth model, producing a first trained artificial
neural network, producing a first output set of synthetic tool responses for the well
logging tool with the first trained artificial neural network, inputting to a second artificial neural network both the input
earth model training set and the first output set of synthetic tool responses for
the well logging tool, producing a second trained artificial neural network
which then serves as the primary trained artificial neural network.
Claim 6
Dependent Claim
129 words
The method of claim 1 further comprising comparing the wellbore logging data to the synthetic tool
responses of the output model to determine amount of mismatch between them, if the mismatch is acceptable, saving the output model for
subsequent use, and if the mismatch is not acceptable, further training the primary
trained artificial neural network producing a secondary trained artificial neural
network and producing therewith a secondary output model that includes a set
of secondary synthetic tool responses, and comparing the wellbore logging data
to the set of secondary synthetic tool responses to determine amount of
mismatch between them, if the mismatch is acceptable, saving the secondary output
model, and if the mismatch is not acceptable again training the secondary
artificial neural network producing a tertiary artificial neural network.
Claim 7
Dependent Claim
13 words
The method of claim 6 wherein acceptable mismatch is no more than
40%.
Claim 8
Dependent Claim
13 words
The method of claim 6 wherein acceptable mismatch is no more than
10%.
Claim 9
Dependent Claim
39 words
The method of claim 6 wherein the primary trained artificial neural
network has a plurality of nodes and the further training of the primary trained
artificial neural network comprises optimizing the nodes of the primary trained
artificial neural network.
Claim 10
Dependent Claim
12 words
The method of claim 9 wherein the nodes are optimized by
backpropagation.
Claim 11
Dependent Claim
138 words
The method of claim 5 further comprising comparing the wellbore logging data to the synthetic tool
responses of a secondary output model produced by the second trained artificial
neural network to determine amount of mismatch between them, if the mismatch is acceptable, saving the secondary output
model, and if the mismatch is not acceptable, further training the second
trained artificial neural network producing a tertiary trained artificial neural
network and producing therewith a tertiary output model that includes a set of
tertiary synthetic tool responses, and comparing the wellbore logging data to
the tertiary set of tool responses to determine amount of mismatch between
them, if the mismatch is acceptable, saving the tertiary output model
for subsequent use, and if the mismatch is not acceptable again training the
tertiary artificial neural network producing a fourth artificial neural network.
Claim 12
Dependent Claim
13 words
The method of claim 11 wherein acceptable mismatch is no more than
255%.
Claim 13
Dependent Claim
13 words
The method of claim 11 wherein acceptable mismatch is no more than
1.5%.
Claim 14
Dependent Claim
39 words
The method of claim 11 wherein the secondary trained artificial neural
network has a plurality of nodes and the further training of the secondary trained
artificial neural network comprises optimizing the nodes of the secondary trained
artificial neural network.
Claim 15
Dependent Claim
13 words
The method of claim 14 wherein the nodes are optimized by back
5propagation.
Claim 16
Dependent Claim
29 words
The method of claim 1 wherein the method is conducted at a well site
having a wellbore extending into the formation, the wellbore logging tool disposed in
the wellbore.
Claim 17
Dependent Claim
21 words
The method of claim 16 further comprising conducting a subsequent
wellbore operation at the well site based on the output model.
Claim 18
Dependent Claim
79 words
The method of claim 17 wherein the subsequent wellbore operation is
from the group consisting of perforating casing lining the wellbore to produce
hydrocarbons from the formation; drilling deeper in the formation; drilling at least one
lateral bore from the wellbore; plugging and abandoning the wellbore; and removing
the wellbore logging tool from the wellbore and then running at least one additional
wellbore tool into the wellbore and using said at least one additional wellbore tool in the wellbore.
Claim 19
Independent Claim
236 words
A method for forward modeling to produce synthetic tool responses for
a well logging tool for an earth formation having a plurality of layers, the wellbore
logging tool in a wellbore in the earth formation and generating wellbore logging
data for a particular part of the earth formation from the group consisting of electrical
resistivity data, acoustic data, vertical seismic profiling data, magnetic data,
gravimetric data, temperature data, and nuclear data, the method comprising producing an input earth model of the particular part of the earth
formation based on the wellbore logging data, inputting the input earth model to a primary trained artificial
neural network, processing the input earth model with the primary trained
artificial neural network to produce an output model that includes a set of
synthetic tool responses for the wellbore logging tool for a plurality of points
in the earth formation, the primary trained artificial neural network produced by
training a first artificial neural network with an input training set producing a
first trained artificial neural network, producing a first output set of synthetic
tool responses for the well logging tool with the first trained artificial neural
network, and inputting to a second artificial neural network both the input
earth model training set and the first output set of synthetic tool responses for
the well logging tool, producing a second trained artificial neural network
which serves as the primary trained artificial neural network.
Claim 20
Dependent Claim
135 words
The method of claim 19 further comprising comparing the wellbore logging data to the synthetic tool
responses of the output model to determine amount of mismatch between them, if the mismatch is no more than 5%, saving the primary trained
artificial neural network for subsequent use, and if the mismatch is not acceptable, further training the primary
trained artificial neural network producing a secondary trained artificial neural
network and producing therewith a secondary output model that includes a set
of secondary synthetic tool responses, and comparing the wellbore logging data
to the secondary set of tool responses to determine amount of mismatch
between them, and if the mismatch is acceptable, saving the secondary output
model, and if the mismatch is not acceptable again training the secondary
artificial neural network producing a tertiary artificial neural network.
References Cited (Backward Citations)
No citation records indexed.
USPTO Annuity & Maintenance Windows
Maintenance schedule is calculated for granted US utility patents.