Back to results
US 6,140,643 AGranted
Method for identification of unknown substances
AssigneeExxonMobil Upstream Research Company
InventorsRoy W. Brown, Chul-Sung Kim
Life of this patent
Estimating the term…
Abstract
The present invention relates to a method for determining the mineral composition of an unknown material using both a fuzzy classification system and a confidence measure of substances identified by the fuzzy classification system. The method can be adapted for identification of pore space and sample points containing more than one material.
Claims19
A method for determining the composition of an unknown substance comprising the steps of:
- 1) creating a database of x-ray spectra from at least one sample of each of a plurality of known substances,
- 2) deriving a fuzzy classification system from said database, said system comprising a plurality of substance membership functions wherein each substance is characterized by its content of each of a plurality of pre-selected elements, said content being defined by an element membership function,
- 3) collecting x-ray spectrum data from one or more analysis points on a sample of said unknown substance,
- 4) selecting for each of said analysis point a candidate substance from said plurality of substance membership functions which most closely matches the x-ray spectrum data of said observation point, and
- 5) classifying each said analysis point as either said candidate substance or an unidentifiable substance using a confidence measure.
2 more elements
A method for determining the composition of an unknown geological sample, said method comprising the steps of:
- 1) creating a training database of normalized x-ray spectra for at least one sample of each of a plurality of known minerals,
- 2) deriving a fuzzy classification system from said database, said system comprising a plurality of substance membership functions wherein each substance is characterized by its content of each of a plurality of pre-selected elements, one of which is carbon, said content being defined by an element membership function,
- 3) collecting x-ray spectrum data, recording measured and calculating normalized values, from a plurality of observation points on a sample of said unknown material, wherein said sample is a geologic sample having a polished surface, prepared by impregnating into the pore space of said sample an epoxy material containing carbon as a signature element,
- 7) selecting a candidate substance for each said observation points from said plurality of substance membership functions which most closely matches the normalized x-ray spectrum data of said observation point,
- 8) classifying each said observation point as either said candidate substance or an unidentifiable material using a confidence measure,
- 9) determining, for each said observation point classified as a candidate substance, whether one or more of the other known substances is also present by comparing measured x-ray intensities of signature elements at each said said observation point to threshold values for each said signature element, and reclassifying observation points determined to contain more than one substance as fractional portions, summing to one, of each substance so identified;
- and
- 10) calculating the overall content of each substance in said sample of said unknown material by dividing the sum of the whole and the fractional amounts of observation points identified as each substance by the total number of observation points identified as any of said known substances.
5 more elements
Dependencies are inferred from the claim text.