Abstract

Background: Madeira Wine can only be sold to the final consumer after a certain aging process that improves its properties. This process requires investment in the wine itself, the storage space and the temperature control and monitorization. The aging process takes a minimum of 3 years but can go up to more than 10 years.

Methods: It has been shown that it can be speed up if the wine is kept under certain conditions with a slightly higher temperature. To control this process constant monitoring is required, namely for the temperature parameter inside the tanks to get a high quality wine. Up to now it is still mostly done in a manual way. To improve the process and reduce costs, a low cost automatic temperature control method has been developed and implement for Madeira Wine, using the tanks of a pilot scale tank.

Conclusion: For this solution a simple Direct Inverse Control presented the best performance.

Keywords: Automatic control, artificial neural network, controller, direct inverse control, models, regression, system identification, wine aging process.

Graphical Abstract