Sizing and control of dynamic energy storage systems toward damping of power swings

This dissertation explores the sizing of shunt connected energy storage devices for damping power system oscillations caused by disturbances in the grid. Damping these low frequency oscillations or power swings is essential because their occurrence over an extended period of time can lead to cascading failures culminating into a blackout. Although many means of control have been explored and developed in the past to quench this swing phenomena, energy storage devices have the advantage of direct monitoring and control of active power to stabilize the power grid. Simulation studies have shown that their optimal placement at certain points in the grid contribute to enhanced stability of the system as a whole. While plentiful of literature makes the case of using energy storage for rotor angle stability, little has been discussed on how to size such devices in terms of their energy rating for desired damping. The research work presented here analyzes this problem by studying the small signal model of a synchronous generator coupled with an energy storage device connected to an infinite bus. This provides a unique insight into the dynamics of the system by showing how the transfer of active power from such a device contributes to the damping of power swings and in turn improves the stability margin of the system. From this, a relation between energy storage size and damping is obtained. The analysis is compared with simulations in PSCAD/EMTDC. Also discussed is the effect of feedback control to tune synchronizing and damping torques of the generator and how it affects energy storage size. This is useful for determining optimal sizes of energy storage for damping purposes. The small signal model is also derived for a case involving two synchronous machines feeding a load and an energy storage device attached to one of them. This model may be the impetus for future studies involving energy storage contribution to stability in a multi-machine scenario and can be extended toward non-linear modeling of such systems.