Automation of gas chromatography instruments. Part II. A knowledge-based system for performance assessment

A knowledge-based system that uses experimentally determined relationships between a gas chromatography (GC) system and the properties of a standard test mixture for automation of quality control and diagnosis of GC instruments is proposed. The approach is to run standard test samples regularly, the results are analyzed automatically, and the system performance is evaluated by the expert-system-based software. When the GC system is set up for standard tests, four Relative Performance Indices are derived that relate the response of this specific instrument to the chemical or physical properties of the samples of interest. Possible component faults are diagnosed in order to ensure that the system operates under perfect condition. This new system can detect faulty components in a GC system and trace the system changes over time. The study indicates that the performance of instruments can be automatically monitored by measuring the changes in the parameters extracted from a standard test sample.