Algorithms for the Qualitative Assessment of Gas Chromatograms

Abstract

An important aspect of automating instruments is automating the detection of system malfunctions that render the generated data unanalyzable. This paper describes a set of algorithms that have been developed to detect and measure features associated with symptoms of analysis failure in gas chromatograms. Each algorithm is individually tested and validated qualitatively by comparing the algorithm output with the opinions of experienced chromatographers. When these algorithms are input into an expert system, the output can be used to detect and diagnose the underlying malfunction that causes the symptoms.