Granted Patent B2

US 12,645,009 · App. 17/644,360 · Granted Jun 2, 2026

Inventors: Hussain Albahrani (Qatif, SA); Arturo Magana-Mora (Dhahran, SA); Mohammad Aljubran (Sayhat, SA); Chinthaka Pasan Gooneratne (Dhahran, SA)

Assignee: SAUDI ARABIAN OIL COMPANY

G01V20/00E21B21/065E21B49/005G06F18/214G06F18/40G06F30/23G06T7/0004G06T7/10G06T7/20G06T7/70G06V10/764G06T2207/10016G06T2207/20081G06T2207/30108

Abstract

Claims (86)

1 . A method, comprising:

taking at least one image of a plurality of returned rock fragments from a wellbore using a camera;

analyzing the at least one image with an image analysis program to detect a caving in the plurality of returned rock fragments;

constructing a model of the caving; and

incorporating the model of the caving into a finite element geomechanics model of the wellbore using a meshing program to create an adjusted model of the wellbore,

wherein the finite element geomechanics model of the wellbore comprises a mesh model of the wellbore,

wherein incorporating comprises:

deriving at least one dimension of the caving from the model of the caving;

removing the at least one dimension of the caving from the mesh model of the wellbore to create an adjusted mesh model of the wellbore; and

executing a finite element geomechanics model program on the adjusted mesh model to create the adjusted model of the wellbore, and

wherein the method enables detection of a wellbore failure in an overburden formation or a reservoir section lacking sufficient logging-while-drilling (LWD) data.

2 . The method of claim 1 , further comprising using the adjusted model of the wellbore to determine drilling window limits for the wellbore, wherein the drilling window limits comprise at least one of a mud weight range and a downhole pressure range to use in the wellbore.

3 . The method of claim 1 , further comprising generating the finite element geomechanics model of the wellbore, the generating comprising:

performing a pre-processing procedure, comprising:

creating an initial mesh of an initial wellbore model, the initial mesh comprising a plurality of elemental units connected by a plurality of nodes;

assigning loads to the initial wellbore model; and

assigning material properties to the initial wellbore model;

inputting outputs from the pre-processing procedure into a finite element geomechanics model program to generate the finite element geomechanics model of the wellbore.

4 . The method of claim 1 , wherein the image analysis program comprises an image segmentation program, wherein the analyzing comprises:

locating boundaries of objects in the at least one image to identify the objects;

determining at least one dimension of the objects; and

identifying the caving based on the at least one dimension.

5 . The method of claim 1 , wherein the at least one image comprises multiple images taken sequentially as the plurality of returned rock fragments are moved across a frame of the camera, and wherein constructing the model of the caving comprises:

tracking the caving at different locations in the frame as the caving moves across the frame in the multiple images;

determining dimensions of the caving from different angles in the different locations in the frame; and

using the dimensions to construct a three-dimensional model of the caving.

6 . The method of claim 1 , wherein the image analysis program comprises a deep-learning program, wherein the analyzing comprises:

manually identifying bounding box coordinates surrounding cavings in a training set of images labeled with the cavings, wherein the training set of images are taken from the at least one image;

training the deep-learning program using the training set of images labeled with cavings; and

after training, using the deep-learning program to identify the caving.

7 . The method of claim 1 , wherein the analyzing the at least one image further comprises identifying a mode of failure of the caving, and wherein the method further comprises:

using the mode of failure to determine a structural state of the wellbore; and

incorporating the structural state of the wellbore into the finite element geomechanics model of the wellbore.

8 . The method of claim 3 , further comprising using the finite element geomechanics model program to:

apply a system of equations to the plurality of elemental units;

apply loads to the initial wellbore model to construct a global stiffness matrix; and

solve the system of equations.

9 . The method of claim 3 , wherein the analyzing the at least one image further comprises using the image analysis program to detect at least one characteristic of the caving, wherein the model of the caving includes the at least one characteristic, and wherein incorporating the model of the caving into the finite element geomechanics model of the wellbore comprises incorporating the at least one characteristic into the material properties assigned to the initial wellbore model.

10 . The method of claim 3 , wherein the analyzing the at least one image further comprises using the image analysis program to identify a rock type of the caving, and wherein generating the finite element geomechanics model of the wellbore further comprises:

retrieving mechanical properties of the rock type from a library of rock properties; and

incorporating the mechanical properties into the material properties assigned to the initial wellbore model.

11 . The method of claim 4 , further comprising using at least one algorithm to classify the objects having the at least one dimension within a first range as cuttings and the objects having the at least one dimension within a different range as cavings.

12 . A method, comprising:

directing drilling fluid returning from a wellbore to a separator;

separating rock fragments from the drilling fluid in the separator;

continuously taking images of the rock fragments separated from the drilling fluid;

sending the images to a computing system;

identifying and characterizing at least one caving from the images using an image analysis program;

using characteristics of the at least one caving identified from the image analysis program to construct a finite element geomechanics model of the wellbore; and

adjusting at least one drilling parameter based on the finite element geomechanics model of the wellbore,

wherein the method enables detection of a wellbore failure in an overburden formation or a reservoir section lacking sufficient logging-while-drilling (LWD) data.

13 . The method of claim 12 , wherein the adjusting the at least one drilling parameter comprises altering an amount of additives to the drilling fluid to change a mud weight in the wellbore.

14 . The method of claim 12 , further comprising continuously updating the finite element geomechanics model of the wellbore each time one of the at least one caving is identified and characterized.

15 . The method of claim 12 , wherein identifying and characterizing the at least one caving comprises:

using a conversion scale that maps pixels in the images to dimensions, wherein the conversion scale defines survey points at an interval of a number of pixels along a selected dimension, and wherein a total amount of the survey points is used to measure the selected dimension; and

using the measured selected dimension to identify the at least one caving.

16 . A system, comprising:

a well system comprising a wellbore extending into a formation;

a cuttings return system, comprising:

a separator fluidly connected to a return line from the wellbore;

a camera positioned proximate the separator; and

a computing system in communication with the camera, the computing system comprising:

an image analysis program comprising instructions for identifying a caving from an image produced by the camera;

a finite element geomechanics model program comprising instructions for generating a finite element geomechanics model of the wellbore; and

a meshing program comprising instructions to:

derive at least one dimension of the caving from the image; and

remove the at least one dimension of the caving from the finite element geomechanics model of the wellbore to create an adjusted model of the wellbore,

wherein the system enables detection of a wellbore failure in an overburden formation or a reservoir section lacking sufficient logging-while-drilling (LWD) data.

17 . The system of claim 16 , wherein the finite element geomechanics model program further comprises:

a pre-processing program comprising instructions for:

creating an initial mesh of an initial wellbore model, the initial mesh comprising a plurality of elemental units connected by a plurality of nodes;

assigning loads to the initial wellbore model; and

assigning material properties to the initial wellbore model; and

a modeling code comprising instructions for:

applying loads to the initial wellbore model to construct a global stiffness matrix; and

solving a system of equations assigned to the plurality of elemental units to reach a convergence close to or equal to equilibrium, wherein the system of equations comprises a minimization of total potential energy equation.

18 . The system of claim 16 , further comprising an additive tank, comprising:

at least one drilling fluid additive; and

a level indicator;

wherein the level indicator is in communication with the computing system.

19 . The system of claim 17 , further comprising:

a library of rock properties comprising mechanical properties of different rock types;

wherein the library of rock properties is digitally stored; and

wherein the finite element geomechanics model program further comprises instructions for:

retrieving the mechanical properties of at least one of the different rock types; and

inputting the mechanical properties into the assigned material properties of the initial wellbore model.

Assignments (1)

Continuity (1)

References Cited (20)