Predictive electricity dispatch through negotiated locational pricing

Joe Warrington, Automatic Control Laboratory, ETH Zurich, Switzerland

ABSTRACT: A predictive mechanism is presented with the aim of reacting to frequent-update predictions of electrical demand and renewable energy generation in competitive electricity markets. The market participants are modelled as price-elastic units, price-inelastic units, and storage  operators. The distributed control algorithm manipulates prices over a time horizon through a negotiation procedure, in order to maximize social welfare while satisfying network constraints. A simple, transparent flow allocation method is used to assign responsibility for constraint violations on the network to individual units, and nodal prices are adjusted  accordingly. Such a framework is appropriate for the inclusion of aggregated household appliances or other 'virtual' market participants realized through smart grid infrastructure. Early success in the pricing algorithm is demonstrated for a densely-populated 39-bus network, where the scheme is shown to allow storage to reduce price volatility in the presence of fluctuating demand.

BIOGRAPHY:
Joe Warrington received the B.A. and M.Eng. degrees from Cambridge University, England in 2008. His masters thesis was on active control of the earth's climate. In the summer of 2007 he studied agent-based collision avoidance algorithms in the Control and Dynamical Systems laboratory at the California Institute of Technology. From 2008-9 he worked on wind turbine control for Wind Technologies Ltd., in Cambridge. He is now a Ph.D. student in the Automatic Control Laboratory at ETH Zurich, Switzerland, where his interests include control on electricity networks and distributed optimization.