SE PhD Prospectus Defense of Elli Ntakou

Starts:
12:00 pm on Friday, January 31, 2014
Ends:
2:00 pm on Friday, January 31, 2014
Location:
15 Saint Mary's Street, Rm 105
TITLE: PRICE DISCOVERY IN DYNAMIC POWER MARKETS WITH LOW-VOLTAGE DISTRIBUTION NETWORK PARTICIPANTS

ABSTRACT: Following a long discussion in the literature originating in Vickrey’s work on dynamic pricing of utility services and its detailed application to Electric Power, dynamic Locational-Marginal-Price (LMP) based Wholesale Power Markets were introduced in England in 1990 and in the US in 1997 resulting in significant productivity dividends: Competitive Power pools were able to accept individual participant bids and offers to clear markets and discover dynamic LMPs that promoted more efficient and reliable service with fewer capacity reserves, located new generation to relieve transmission congestion and to lower supply cost to consumers, and more.
Nevertheless, mature, yet unexploited opportunities can be sought in the potential of extensive load-side market participation and the use of Distribution network Locational Marginal Prices (DLMP). In particular, we note that distribution network costs, accounting for as much as 35% of low voltage power costs, are priced today at their average cost. We propose space and time sensitive pricing reflecting marginal costs of (i) ancillary services, such as distribution network reactive power compensation and voltage control, (ii) marginal distribution network line losses and (iii) distribution asset congestion and life degradation. Present distribution network average pricing practice deprives millions of consumers from the opportunity to match their preferences to distribution system marginal costs; it wastes the opportunity to capture significant cost reducing efficiencies and to assist the cost efficient integration of clean distributed generation.
Critical developments that have occurred since wholesale power markets made their debut- affordable communication, computation, sensing, actuation, and the advent of flexible loads and ubiquitous power electronics- advocate a major power market reform. The potential of smart appliance demand response has been recognized and studied extensively under DOE funding. However, most research till now has focused on direct control or centrally broadcasted Utility signals directed to various smart appliance types.
My research is complementary in scope to Pacific Northwest National Laboratory research. It focuses on the extension of Locational Marginal Price principles to (i) incorporate marginal costs of real and reactive power, line losses, voltage control, and distribution asset life degradation, and (ii) enable medium and low voltage consumers and distributed generation to provide reactive power compensation, voltage control and line loss reduction. We aspire to contribute to the debate on whether detailed DLMP-based markets are worth considering. To this end, we propose an explicit market clearing framework and apply it to a realistic distribution network feeder adopted from Southern California Edison data. We solve for day-ahead market clearing prices and quantities for real and reactive power consumed or produced at each load or generation point in the network so as to optimize distribution utility cost minus distributed participant utility subject to full AC load flow relations and voltage magnitude constraints. Distribution network marginal cost based decisions and pricing inroads have also been reported by other researchers. We, however, generalize proposing a complete market framework that (i) includes distribution utility and distributed participant costs and benefits and (ii) models degrees of freedom ranging from the ability to delay and reschedule consumption to putting excess power electronics resources to dual use for VAr compensation.
This work addresses issues like: (i) the non-convexity of load flow equality constraints, (ii) computational complexity, (iii) communication requirements, (iv) practical distribution feeders, their assets, capabilities and performance, (v) apparent differences among residential, light industrial, and commercial distribution feeders, (vi) the resulting clearing prices and the distribution network rent that they generate, (vii) location and income of distributed generation, (viii) reactive power market size, and (ix) proximal message passing distributed decision solutions.

COMMITTEE:: Advisor: Michael Caramanis, SE/ME; Ioannis Paschalidis, SE/ECE; Pablo Ruiz, SE/ME; Ali Abur, by Special Service Appointment, Northeastern University