Exploiting intelligent systems techniques within an autonomous regional active network management system

Abstract

This paper discusses AuRA-NMS, an autonomous regional active network management system currently being developed in the UK through a partnership between several UK universities, two distribution network operators (DNO) and ABB. The scope of control to be undertaken by AuRA-NMS includes: automatic restoration, voltage control, power flow management and implementation of network performance optimisation strategies. Part of the scientific aims of the AuRA-NMS programme is the investigation and comparison of the use of different techniques for making the control decisions above. The techniques under consideration range from the use of optimisation techniques, such as OPF, to the use of intelligent systems techniques, such as constraint programming, case-based reasoning and AI planning. In this paper the authors consider the role that intelligent systems techniques could play within active network management and reports on preliminary results gathered from the testing of prototype software running on commercially available IEC 61850 compliant substation computing hardware, connected to a real time power systems simulator. The importance of an appropriate comparative testing methodology, as well as the need for assessing the robustness of different techniques in the presence of communication failures and measurement errors, is also discussed. A key element in the development of AuRA-NMS is the use of multi-agent systems (MAS) technology to provide a flexible and extensible software architecture in which the techniques above can be deployed. As a result, the use (MAS) in conjunction with IEC 61850 and the common information model within AuRA-NMS is described.