Smart Grid

Leader

Dr. Dennis Jarvis, School of Computing Sciences

Members

• Prof. Qing-Long Han, School of Computing Sciences
• Dr. Jacqueline Jarvis, School of Management and Information Systems
• Dr. Amanullah Maung Than Oo (College of Engineering & Built Environment)
• Dr. Alison Payne (Post Doctoral Research Fellow, CINS)
• Dr. Yang Xiang, School of Management and Information Systems

Activities

The smart grid is a broad collection of technologies that delivers an electricity network that is flexible, accessible, reliable and economic. Smart grid facilitates the desired actions of its users and these may include distributed generation, the deployment of demand management and energy storage systems or the optimal expansion and management of grid assets. Governments around the world are now recognising  the urgency to begin the evolution of existing power grids to smart grids [1].

From a research perspective, the key issue with the smart grid is not its "smartness" per se. Rather it is the interplay that arises between consumers, providers, generators on the one hand and a range of autonomous and semi-autonomous systems on the other. Furthermore, parts of that interplay will take place over public networks, so bandwidth constraints and both Quality of Service and security guarantees all need to be factored in. What is required is first of all a coherent software architecture that can support a spectrum of behaviour ranging from autonomous line protection in sub-stations through to load balancing by providers and usage optimisation by consumers. An agent-based architecture provides the obvious solution, but based on our experience in both the manufacturing and defence sectors, we would expect that the major challenge will be in the specification of how teams of agents (including humans) should work in concert to achieve system goals and not in the provision of intelligent behaviour for individual entities.

Traditional multi-agent architectures approaches provide limited, if any, support for the explicit specification of team behaviour, instead delegating responsibility for team formation and task execution to the individual entities concerned. While such delegation is an important aspect of team behaviour, predefined behaviour patterns are also important, especially in situations where deterministic behaviour at the system level is required, as in manufacturing, business and defence. CINS researchers have pioneered this explicit modelling of team behaviour through their involvement in the development and deployment of the first commercial team modelling framework (JACK Teams) and then a next generation successor (GORITE). The intent is that GORITE will provide the architectural underpinning for our smart grid research program.

The availability of GORITE gives us the opportunity to explore a number of niche research areas in an integrated manner - distributed control, security and information systems. These opportunities are summarised below:

[1] Marris, E., 2008, "Upgrading the Grid", Nature, Vol. 454, July 31, pp. 570-573