Reliability of operating window identified from process data

Heimo Ihalainen, Risto Ritala, Olli Saarela

    Research output: Chapter in Book/Report/Conference proceedingConference article in proceedingsScientificpeer-review


    This paper discusses data-based operating windows as a tool for process management and development. In particular identification of the operating window and its uncertainty are analyzed. The operating window is determined by maximizing either the mutual information (static) or entropy transfer (dynamic). An industrial example shows that the entropy of the indicator variable is reduced to half by an operating window specified with only few variables, selected amongst over 3000 candidates. Test model based simulations suggest that such few-variable operating windows can be reliably identified from datasets having lengths of a few thousand observations.
    Original languageEnglish
    Title of host publication23rd European Symposium on Computer Aided Process Engineering
    EditorsAndrzej Kraslawski, Ilkka Turunen
    ISBN (Print)978-0-444-63234-0
    Publication statusPublished - 2013
    MoE publication typeA4 Article in a conference publication

    Publication series

    SeriesComputer Aided Chemical Engineering


    • mutual information
    • operating window
    • process management
    • transfer entropy
    • uncertainty


    Dive into the research topics of 'Reliability of operating window identified from process data'. Together they form a unique fingerprint.

    Cite this