Massive coordination of dispersed generation using PowerMatcher based software agents
暂无分享,去创建一个
One of the outcomes of the EU-Fifth framework CRISPproject [ http://crisp.ecn.nl ], has been the development of a real -time control strategy based on the application of distributed intelligence (ICT) to coordinate demand and supply in electricity grids. This PowerMatcher [1] approach has been validated in two real-life and real-time field tests. The experiments aimed at controlled coordination of dispersed electricity suppliers (DG-RES) and demanders in distributed grids enabled by ICT networks. Optimization objectives for the technology in the tests were minimization of imbalance in a commercial portfolio and mitigation of strong load variations in a distribution network with residential micro-CHPs. With respect to the number of ICT-nodes, the field tests were on a relatively small-scale. However, application of the technology has yielded some very encouraging results [2][3] in both occasions. In the present paper, lessons learned from the field experiments are discussed. Furthermore, it contains an account of the roadmap for scaling up these field-tests with a larger of number nodes and with more diverse appliance/installation types. Due to its autonomous decision making agent-paradigm, the PowerMatcher software technology is expected to be widely more scaleable than central coordination approaches. Indeed, it is based on microeconomic theory and is expected to work best if it is applied on a massive scale in transparent market settings. A set of various types of supply and demand appliances was defined and implemented in a PowerMatcher software simulation environment. A massive amount of these PowerMatcher node-agents each representing such a devicetype was utilized in a number of scenario calculations. As the production of DG -RESresources and the demand profiles are strongly dependent on the time-of-year, climate scenarios leading to operational snapshots of the cluster were taken for a number of representative periods. The results of these larger scale simulations as well as scalability issues, encountered, are discussed. Further issues covered are the stability of the system as reflected by the internal price development pattern that acts as an ‘invisible hand’ to reach the common optimisation goal. Finally, the effects of scaling -up the technology are discussed in terms of possible ‘emergent behaviour’ of subsets in the cluster and primary process quality of appliances operating concertedly using the PowerMatcher.
[1] J. K. Kok,et al. Wind Turbines and Heat Pumps - Balancing wind power fluctuations using flexible demand , 2006 .
[2] W Wim Zeiler,et al. Learning about smart systems for comfort management and energy saving in office buildings , 2003 .
[3] J. K. Kok,et al. The PowerMatcher: Multiagent Control of Electricity Demand and Supply , 2006 .
[4] J. K. Kok,et al. Field tests applying multi-agent technology for distributed control: virtual power plants and wind energy , 2006 .