NUMERICAL METHOD OF ESTIMATING PHYSICAL PROPERTIES OF THREE-DIMENSIONAL POROUS MEDIA

Patent №

US 6,516,080

Granted

2003-02-04

Filed 2000

Owner

BOARD OF TRUSTEES OF THE LELAND STANFORD JUNIOR UNIVERSITY, THE

AI components

1

vision

Assignment

Recorded

Dataset

AIPD

2023_r1 edition

Application

09543829

A section is prepared from a specimen of porous media, e.g., Rock, typically including filling the pore spaces with a dyed epoxy resin. A color micrograph of the section is digitized and converted to an n-ary index image, for example a binary index image. Statistical functions are derived from the two-dimensional image and used to generate three-dimensional representations of the medium. Boundaries can be unconditional or conditioned to the two-dimensional n-ary index image. Desired physical property values are estimated by performing numerical simulations on the three-dimensional representations. For example, permeability is estimated by using a Lattice-Boltzmann flow simulation. Typically multiple equiprobable three-dimensional representations are generated for each n-ary index image, and the multiple estimated physical property values are averaged to provide result.

VisionG06T 17/00G06T 7/40G06T 7/97

AI classification

Vision1.00
Planning0.16
Machine learning0.01
Evolutionary computation0.00
Natural language0.00
Knowledge representation0.00
AI hardware0.00
Speech0.00

Ownership

BOARD OF TRUSTEES OF THE LELAND STANFORD JUNIOR UNIVERSITY, THE

assignment · 106920597

Assignors

NUR, AMOS M.

On an employer assignment, the assignors are typically the inventors.

© 2026 NYSGPT2525 LLC