Data Analytics and Monitoring of the Tennessee Eastman Process

Vast amounts of data are generated in process facilities from sensors that record information from all aspects of the process including flows, temperatures, and levels. On the other hand, Product quality measurements are necessary to make sure that the process produces on-spec products and operates efficiently in material and energy consumptions. The Tennessee Eastman process was established as a challenging problem for researchers to test new ideas for process monitoring, optimization, and control. The tasks include building a PCA model based on the normal case data, and establish T2 and Q control limits quality data, CCA using quality data, concurrent CCA monitoring using quality data, and analysis of quality-relevant and irrelevant Cases.