Patent · US Active

Multi-channel machine learning model-based inversion

US11914096B2 · kind B2 · utility

0Cited by
3References
20Claims
0Family size

Assignee

Inventors

Key dates

Filing dateJan 3, 2022
Grant dateFeb 27, 2024
Priority date
Expiry dateJan 3, 2042

Classification

  • Technology area (CPC G)Physics
  • CPC primaryG06N20/00
  • WIPO fieldComputer technology
  • WIPO sectorElectrical engineering

Abstract

A method for identifying a collar using machine learning may include acquiring one or more measurements from one or more depth points within a wellbore including a tubular string, training a machine learning model using a training dataset to create a trained machine learning model, and identifying at least one hyperparameter using the trained machine learning model. The method may further include creating a synthetic model, wherein the synthetic model is defined by one or more pipe attributes, minimizing a mismatch between the one or more measurements and the synthetic model utilizing the at least one hyperparameter, updating the synthetic model to form an updated synthetic model, and repeating the minimizing the mismatch with the updated synthetic model until a threshold is met.

Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.