Predicting ethanol concentration behavior of future harvests using Knowledge Discovery in Database

Abstract

The growing demand for higher productivity in the sugar-alcohol sector has required higher levels of automation in new and existing production plants. However, with increasing the number of sensors and equipment emerges a correspondent growth in the amount of data generated. Although the industry stores most of these data in dedicated data warehouse they are rarely used in future analysis due to the inherent technological challenge to properly cope with the large amount of data. This paper proposes the usage of a Knowledge Discovery in Database (KDD) process as a powerful tool to assist one in obtaining relevant industrial behavior from the stored data with the purpose of allowing quality and efficiency analysis. The experiments conducted with data collected in an industrial sugarcane plant successfully demonstrate that it is possible to apply the KDD to predict the ethanol concentration of future harvests.