Method to recommend design practices that increase the probability of meeting cementing job objectives

Patent metadata

Field	Value
Publication number	US-12406114-B2
Application number	US-202117409303-A
Country	US
Kind code	B2
Filing date	Aug 23, 2021
Priority date	Aug 23, 2021
Publication date	Sep 2, 2025
Grant date	Sep 2, 2025

Abstract

A method of designing a cement pumping procedure of a wellbore isolation barrier includes using a design application that receives customer inputs, which include wellbore data and a job objective. The design application can retrieve a cement pumping procedure from a data source comprising a series of sequential steps to achieve the job objective. The design application can load the customer inputs into the cement pumping procedure, access a database of best practices, calculate a probability score for achieving the job objective based on a model, and recommend modifying one or more steps of the cement pumping procedure with one or more best practices in response to the probability score being below a threshold.

First claim

What is claimed is:

1. A method of cementing a wellbore penetrating a subterranean formation, comprising:
receiving one or more customer inputs, by a design application executing on a computer system;
initializing a design cement placement procedure, by the design application, by applying the customer inputs into a generic cement placement procedure;
calculating, by the design application, a bill of materials and a cement equipment list from the design cement placement procedure, wherein the design application modifies at least one generic methodology to at least one design methodology with the bill of materials;
generating, by a machine learning process, a final cement placement procedure by modifying the design cement placement procedure with one or more prescribed methodologies in response to a probability value for the design cement placement procedure achieving a job objective with the one or more prescribed methodologies being greater than the probability value for the design cement placement procedure achieving a job objective without the one or more prescribed methodologies; and
placing cement in the wellbore in accordance with the final cement placement procedure, wherein the placing of the cement in the wellbore comprises:
transporting a cement blend and cement pumping equipment to a well site, wherein the cement blend is included in the final cement placement procedure;
retrieving, by a managing application executing on the computer system, the final cement placement procedure;
beginning the final cement placement procedure by the managing application; and
continuing the final cement placement procedure, by the managing application, in response to the probability value being above a threshold value for achieving the job objective.

2. The method of claim 1 , further comprising:
retrieving the generic cement placement procedure, by the design application, from a database, wherein the generic cement placement procedure is a cement placement procedure comprising a series of sequential stages to achieve the job objective, wherein the stages comprise one or more steps with the at least one generic methodology, and wherein the stages and the at least one generic methodology are machine learning inputs.

3. The method of claim 2 , further comprising:
modifying, by the design application, the design cement placement procedure by inputting the bill of materials and one or more customer inputs into one or more wellbore construction models, wherein the wellbore construction models comprise a casing design model, casing stretch model, cement strength model, well control model, formation strength model, or dynamic loading model, wherein at least one wellbore construction model recommends modifying the bill of materials in response to a calculated design value being lower than a design threshold.

4. The method of claim 3 , wherein:
a machine learning process classifier identifies the at least one design methodology of the at least one stage of the design cement placement procedure;
the machine learning process classifier determines a methodology grade for the design methodology within the database;
the machine learning process classifier identifies a corresponding prescribed methodology within the database;
wherein the machine learning process compares the methodology grade of the design methodology to the methodology grade of the corresponding prescribed methodology; and
wherein the machine learning process retrieves the prescribed methodology in response to the prescribed methodology having a higher methodology grade.

5. The method of claim 4 , further comprising:
comparing, by the machine learning process, a first probability value for achieving a stage objective of at least one stage of the design cement placement procedure with at least one design methodology within at least one step of the stage to a second probability value for achieving the stage objective with a prescribed methodology, wherein the prescribed methodology is retrieved from the database, and wherein the prescribed methodology has a higher methodology grade than the design methodology; and
replacing, by the machine learning process, the at least one design methodology within the at least one step of the stage with the prescribed methodology in response to the second probability value being greater than the first probability value for achieving the stage objective.

6. The method of claim 1 , wherein:
the one or more customer inputs comprise a plurality of wellbore data, a plurality of customer design inputs, and a job objective, and the customer design inputs comprise a bill of materials and one or more construction plans.

7. The method of claim 6 , wherein:
the wellbore data comprises a wellbore location, a wellbore history, and a plurality of sensor data, wherein the wellbore location comprises well names, a lease location, global positioning satellite (GPS) coordinates, an internal designation, or a combination thereof;
the wellbore history comprises a wellbore survey, a wellbore drilling path, a wellbore production fluid analysis, a wellbore drilling fluid, and a wellbore construction history; and
the plurality of sensor data comprises well logging, wellbore production sample, and well control data.

8. The method of claim 6 , wherein:
the job objective comprises wellbore isolation, a location of top of cement, a kick off plug, a shoe test, or a combination thereof.

9. The method of claim 1 , further comprising:
mixing a cement slurry, by the cement pumping equipment, according to the final cement placement procedure; and
pumping the cement slurry according to the final cement placement procedure.

10. The method of claim 1 , further comprising:
transporting a downhole tool to a well site, wherein the downhole tool is included in the final cement placement procedure; and
coupling the downhole tool to a casing according to the final cement placement procedure.

11. A method of cementing a wellbore, comprising:
retrieving, by a machine learning process executing on a computer system, a completed cement placement procedure;
identifying, by a machine learning classifier of the machine learning process, a format of the completed cement placement procedure, wherein the format comprises a job objective, a plurality of stages, a plurality of wellbore data, and a plurality of measured field data;
retrieving, by the machine learning classifier, the plurality of stages from the completed cement placement procedure;
producing, by the machine learning process, a combined methodology grade for each completed stage by comparing at least one completed stage of the plurality of stages to the plurality of measured field data corresponding to the completed stage, wherein the combined methodology grade comprises a plurality of completed methodologies within the completed stage;
validating, by the machine learning process, the combined methodology grade by comparing a predictive stage objective grade using the completed methodology grade to the completed stage objective grade to determine an error value, wherein the completed stage objective grade comprises the plurality of measured field data, a completed job objective grade, and a plurality of completed stage objective grades;
training the machine learning process to reduce the error value;
generating, by the machine learning process, a final cement placement procedure; and
placing cement in a wellbore in accordance with the final cement placement procedure, wherein the placing of the cement in the wellbore comprises:
transporting a cement blend and cement pumping equipment to a well site for a job, wherein the cement blend is included in the final cem

Assignees

• Halliburton Energy Services Inc

Inventors

• Jandhyala Siva Rama Krishna
• Singh John Paul Bir
• Yerubandi Krishna Babu
• Vargo Jr Richard Frank

Classifications

• E21B2200/22

  Fuzzy logic, artificial intelligence, neural networks or the like · CPC title

• E21B2200/20

  Computer models or simulations, e.g. for reservoirs under production, drill bits · CPC title

• G06N20/00

  Machine learning · CPC title

• G06F30/27 Primary

  using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model · CPC title

• E21B33/14 Primary

  for cementing casings into boreholes · CPC title