Fisher information based time-series segmentation of streaming process data for monitoring and supporting on-line parameter estimation in energy systems

László Dobos, János Abonyi

Research output: Contribution to journalArticle


Advanced process engineering techniques require proper process model. This motivates the process developers to find those input-output time series segments, which possess the highest information content in the identification process. In this study a novel streaming multivariate time-series segmentation method is going to be introduced. This is based on the Fisher information matrix and for supporting to segregate information-rich time-series segments. The Krzanowski similarity measure is applied to determine the similarity of the segments in the implemented segmentation process, which is able to handle streaming process data. To prove the efficiency of the proposed methodology, a case study of a district heating network is examined in details. The task is to separate the time series segments with high information content in the parameter estimation point of view.

Original languageEnglish
Pages (from-to)1844-1848
Number of pages5
JournalComputer Aided Chemical Engineering
Publication statusPublished - Jun 20 2011



  • District heating network
  • Fisher information matrix
  • Streaming data
  • Time-series segmentation

ASJC Scopus subject areas

  • Chemical Engineering(all)
  • Computer Science Applications

Cite this