| 摘要 |
The method and apparatus are used to determine class, grade and properties of fuel samples, regardless of ambient, instrument, or sample temperature, using mathematical correlations between fuel class, grade and properties and their spectra developed from a database of samples with measured properties and spectra. The ability to measure a fuel sample using the present method and apparatus is useful in identifying unknown fuel samples, determining suitability in equipment, and monitoring and controlling fuel processes, such as blending operations, distillation, and synthesis. |
| 主权项 |
1. A method for determining spectroscopically the value of at least one property of a fuel sample of unknown character, comprising:
providing a coarse mathematical model, created using a database, in which either or both of (a) the nature and values of characteristics that are indicative of fuels of at least two classes, or (b) one or more spectral representations of the nature and values of characteristics that are indicative of fuels of said at least two classes, taken cumulatively, at temperatures over a range of common values, are correlated to measured spectra in a selected spectral range, said fuel characteristics including, directly or indirectly, chemical functional groups and a plurality of properties; providing a multiplicity of fine mathematical models, created independently using said database, in each of which a plurality of spectral components are correlated to a property of fuels within each of said at least two classes, such correlations being based upon either or both of (a) the nature and value of fuel characteristics, or (b) one or more spectral representations of such characteristics, taken cumulatively and at temperatures over said range of common values, separate pluralities of said multiplicity of fine mathematical models being grouped together so as to provide at least a first multiplicity of fine model groups, a first plurality of said first multiplicity of fine model groups being peculiar to fuels within at least a first one of said two classes, and a second plurality of said first multiplicity of fine model groups being peculiar to fuels within at least a second one of said two classes; obtaining, through measurements, a spectrum, within said spectral range, of a fuel sample of unknown character, at least at one temperature within said range of common values; performing a coarse analysis of said obtained spectrum by comparing said obtained spectrum to said measured spectra, using said coarse mathematical model, so as to identify a class to which the fuel of said sample belongs; and performing a fine analysis of said obtained spectrum by comparing said obtained spectrum to said measured spectra, using said first plurality of fine model groups if said class to which said fuel sample belongs is determined, by said course analysis, to be said first one of said two classes, or using said second plurality of fine model groups if said class to which said fuel sample belongs is determined, by said coarse analysis, to be said second one of said two classes, so as to thereby determine a value for said correlated property of said fuel of said sample. |
