Finding the composition of gas mixtures by a phthalocyanine-coated QCM sensor array and an artificial neural network

This paper presents a system, which is made of an array of eight phthalocyanine-coated QCM sensors and an ANN to find the corresponding composition of a gas mixture. The digital data collected from the sensor responses were preprocessed by a sliding window algorithm, and then used to train a three layer ANN to determine the gas compositions. The system is tested with the following gas mixtures: (1) ethanol–acetone, (2) ethanol–trichloroethylene, (3) acetone–trichloroethylene. The success rate of the system in identifying the constituent component amounts is 84.5 and 94.3%. Similarly, overall average prediction error is 10.6%.