Title: Graphical tools for generating qualitative redundant relations

Authors: Rahma Smaili; Rafika El Harabi; Mohamed Naceur Abdelkrim

Addresses: Research Unit MACS, National Engineering School of Gabès, University of Gabès, Omar Ibn Elkhatab Street, 6029 Gabès, Tunisia ' Research Unit MACS, National Engineering School of Gabès, University of Gabès, Omar Ibn Elkhatab Street, 6029 Gabès, Tunisia ' Research Unit MACS, National Engineering School of Gabès, University of Gabès, Omar Ibn Elkhatab Street, 6029 Gabès, Tunisia

Abstract: To perform fault diagnosis for multi-energy systems, the causal graph-based methods are well-suited for modelling and monitoring purposes. The present paper is concerned with the merging of the quantitative model (bond graph) and the qualitative reasoning (signed directed graph) and its subsequent is introduced as a new causal graph framework. The integrated model can be involved to generate qualitative fault signatures matrix deduced from the qualitative redundant relations. Compared with the classical bond graph-based monitoring, the proposed mechanism can reduce the number of redundant relations which depend on the qualitative state of known variables and enhance the process monitoring in normal situations as well as in the presence of failures without any need of numerical calculations. The performance of the developed approach is explained using two illustrative examples (systems with single and coupled energies) and discussed in detail with the existing bond graph method.

Keywords: signed directed graph; bond graph; quantitative reasoning; qualitative reasoning; fault monitoring; redundant relations optimisation; fault diagnosis; failures; process monitoring.

DOI: 10.1504/IJRS.2016.076344

International Journal of Reliability and Safety, 2016 Vol.10 No.1, pp.22 - 47

Received: 25 May 2015
Accepted: 30 Dec 2015
Published online: 04 May 2016 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article