Reliability implications of price responsive demand : a study of New England's power system

Abstract

With restructuring of the traditional, vertically integrated electricity industry come new opportunities for electricity demand to actively participate in electricity markets. Traditional definitions of power system reliability have carried over from the vertically integrated market structure, but this thinking will become increasingly problematic as the proportion of electricity demand responsive to market prices increases. Using New England as an example, this thesis highlights these difficulties by employing a Probabilistic Production Costing model modified to account for price responsive demand. A Neural Gas clustering algorithm is used to deal with the time-varying nature of price responsive demand. We show that neglecting to account for price responsive demand could result in underestimating system reliability by traditional measures, and discuss a possible new metric to help the transition to thinking of reliability as one aspect of whole market performance.