TY - JOUR T1 - Automatic Sensor System for the Continuous Analysis of the Evolution of Wine JF - American Journal of Enology and Viticulture JO - Am. J. Enol. Vitic. DO - 10.5344/ajev.2014.14103 SP - ajev.2014.14103 AU - Jesús Lozano AU - José Pedro Santos AU - José Ignacio Suárez AU - Mariano Cabellos AU - Teresa Arroyo AU - Carmen Horrillo Y1 - 2014/10/30 UR - http://www.ajevonline.org/content/early/2014/10/22/ajev.2014.14103.abstract N2 - The development of an in situ and on-line electronic nose installed in a wine cellar for the continuous measurement of the wine evolution, is presented in this paper. The system has a novel sampling method that extracts the aroma directly from the tanks where wine is stored; and it automatically carries the volatile compounds to the sensor cell. A tin oxide multisensor, prepared by RF sputtering onto an alumina substrate and doped with chromium and indium, is used. The whole system is fully automated and controlled by computer and can be supervised on the Internet. Linear techniques as principal component analysis (PCA) and nonlinear ones as Artificial Neural Networks (ANN) are used for pattern recognition, and Partial Least Squares (PLS) is used for predicting GC-MS analysis. Results show that system can detect the evolution of two different wines along 9 months stored in tanks. Correlation coefficients near to 1 are obtained in the prediction of several volatile organic compounds (VOCs), i.e. ethyl butyrate, isobutyric acid, isobutyl acetate, hexyl acetate and ethyl octanoate. This system could be trained for monitoring wine preservation and evolution in tanks and therefore detecting off-odours of wine and warning the wine expert to correct it as soon as possible, preventing the wine spoilage and improving its final quality. ER -