Kalinichenko A. Intelligent multisensor system for identification and quality assessment of food products

Thesis is devoted to the development of an intelligent multisensor system with electronic nose methodology, including the study of obtaining optimal arrays of quartz crystal microbalance sensors with cross-sensitivity and good discriminating ability and analysis of the multivariate response of the system using machine learning methods. Proposed coating materials of sensors based on sorbents with different polarity, stability and sensibility (poly(ethylene glycol sebacate), poly(ethylene glycol adipate), dicyclohexano-18-crown-6, Triton X-100, poly(ethylene glycol) 2,000, poly(ethylene glycol succinate), poly(diethylene glycol succinate), polyvinylpyrrolidone, Tween 80, trioctylphosphine oxide, beeswax), which are characterized by uniform, mesoporous and rough coatings structure, and reproducibility of sorption properties. The high sorption capacity of some sensors and the variation of selectivity in arrays to the main volatile markers and classes of food VOCs allowed to extract their identification features in a multicomponent gas sample during the identification, assessment of quality and safety of objects. An intelligent chemical recognition system has been developed to solve the qualitative and quantitative analysis tasks of objects in one measurement, which includes recommendations for the construction of feature space, synthesis of neural networks with different architectures (PNN, LVQ and FFNN) and optimization of learning procedures to build classification and regression models based on multidimensional electronic nose data. The effectiveness of the approaches is confirmed by the developed methods of rapid assessment of food authenticity, detection and quantitative assessment the mass content of soy protein in sausages, microbial counts prediction (QMAFAnM) in meat and sausage products, peroxide value of vegetable oils, creation of intelligent portable electronic nose that can be used for express analysis of gases in the food and chemical industries, medicine and for environmental monitoring.