Mining redundant industrial alarm occurrences with association rules extraction and complex networks modeling
Issue title: Special Supplement Issue in Section A and B: Selected Papers from the ISCA International Conference on Software Engineering and Data Engineering, 2010
Affiliations: Department of Computer Science, Universidade Federal de Minas Gerais – UFMG, Minas Gerais-MG, Brazil | Department of Computer Science, Winona State University, Winona, MN, USA
Abstract: Alarm management is a research area that is growing rapidly on industrial automation. One of the major difficulties in alarm rationalization, in which the volume of generated alarms is reduced to an appropriate number so that a human being can handle them, is to identify patterns that might indicate unnecessary alarms in the middle of files and databases containing tens of thousands of daily records. This work presents a new approach to analyze alarm occurrences, combining several techniques, such as: sequence mining, association rules extraction with MNR (Minimum Non Redundant Association Rules), cross-correlation analysis, and complex network modeling for visualization. The combination of different techniques creates a more comprehensive alternative to the detection process. The solution's performance, in terms of accuracy, shows improvements over the current approaches, resulting in a more reliable and predictable alternative for identification of meaningful patterns.
