Abstract
Bioethanol production faces control challenges due to its biological alive-process nature and due to the scaling up from the well characterized environment of laboratories to the less
controlled industrial facilities. Although second generation technologies -aiming to process lignocellulosic feedstock- are already approaching the market phase, much progress has been
done in the first generation technologies, based on starch sacarification. Nonetheless, still room for improvement is left to exploit first generation knowledge embedded in the data gathered
during years of continuous operation. In this paper ongoing research for an extensive analysis of such operational data and the possibilities lying in its modeling using Artificial Intelligence
(AI) techniques to better explain deviations in the performance of the process is shown. Preliminary results show great possibilities to enlighten the still-grey areas in starch fermentation, while paving the way to extensive application also on second generation technologies.
