Genetic algorithms applied to pattern recognition analysis of high-speed gas chromatograms of aviation turbine fuels using an integrated Jet-A/JP-8 database

High-speed gas chromatography was used to develop a potential method to type civilian and military jet fuels. A database of 212 gas chromatograms of neat jet fuel samples representing common aviation turbine fuels found in the United States (Jet-A, JP-5, JP-7, JP-8, and JPTS) was mined using a genetic algorithm, which was necessary because of the similarities of the gas chromatograms in the database. Principal-component models developed from gas chromatography peaks identified by the genetic algorithm were able to correctly classify the gas chromatograms of neat jet fuels, and these models were also able to successfully classify the gas chromatograms of jet fuels that had undergone weathering in a subsurface environment. The present study, which is a logical extension of an earlier effort was undertaken because of the change from JP-4 to JP-8 as the principal U.S. Air Force fuel. (C) 1999 Academic Press.