Application of Dynamic Generation Control for Predatory Competitive Advantage in Electric Power Markets

Abstract

This work extends investigations into the potential for coordinated control of a group of generators to act in a manner that destabilizes other machines, while maintaining nearly satisfactory performance within this control group. This control scheme raises the specter that dynamic control of generators might be used to gain predatory competitive advantage. To assess the risk of this behavior, it is useful to more carefully examine its technical feasibility. The work to be presented here will examine two questions in this context. First, to what extent is this control robust with respect to inclusion of realistic dynamic models, including the effect imprecisely known data describing the network and competitive generator units? Second, to what extent can this control be implemented using only local measurement information, such as that available from generator terminal measurements? This work demonstrates that the addition of governor and turbine dynamics into the model does not impede the ability of the feedback control to act in the predatory manner. It also demonstrates the possibility that the system state can be adequately recreated from a localized measurement set, such that this estimated state can act as a feedback source. Moreover, for the case study examined, the combined state observer and feedback design are shown to be quite robust, even when the design is based upon a model having parameter errors on the order of 20%.