Discovery of behavior in industrial plants: A KDD based proposal

Nowadays, with the increasing technological computational advances in industrial environments, the amount of data monitored and stored in the industry is considered extensive and continuous. However, not all of the stored variables are analyzed, considering that there is a great difficulty in knowing which variables are useful and which variables should be analyzed. This paper proposes the use of Knowledge Discovery in Database (KDD) process as a powerful tool to predict the behavior of industrial processes, more specifically, a sugar-ethanol production plant. The tests were conducted with data obtained from a level control of a didactic industrial plant. The obtained experimental results showed that it is possible to apply the KDD process in industrial environments, in such a way that one may identify types of behaviors inherent in industrial processes.