Issues in integrating existing multi-agent systems for power engineering applications

Multi-agent systems (MAS) have proven to be an effective platform for diagnostic and condition monitoring applications in the power industry. For example, a multi-agent system architecture, entitled condition monitoring multi-agent system (COMMAS) (McArthur et al., 2004), has been applied to the ultra high frequency (UHF) monitoring of partial discharge activity inside transformers. Additionally, a multi-agent system, entitled protection engineering diagnostic agents (PEDA) (Hossack et al., 2003), has demonstrated the use of MAS technology for automated and enhanced post-fault analysis of power systems disturbances based on SCADA and digital fault recorder (DFR) data. In this paper, the authors propose the integration of COMMAS and PEDA as a means of offering enhanced decision support to engineers tasked with managing transformer assets. By providing automatically interpreted data related to condition monitoring and power system disturbances, the proposed integrated system offer engineers a more comprehensive picture of the health of a given transformer. Defects and deterioration in performance can be correlated with the operating conditions it experiences. The integration of COMMAS and PEDA has highlighted the issues inherent to the inter-operation of existing multi-agent systems and, in particular, the issues surrounding the use of differing ontologies. The authors believe that these issues need to be addressed if there is to be widespread deployment of MAS technology within the power industry. This paper presents research undertaken to integrate the two MAS and to deal with ontology issues.