Real-time Integration of Energy Storage

Increasing dynamics in power systems on account of renewable integration, electric vehicle penetration and rising demands have resulted in the exploration of energy storage for potential solutions. Recent technologyand industry-driven developments have led to a drastic decrease in costs of these storages, further advocating their usage. This thesis compiles the author’s research on optimal integration of energy storage. Unpredictability is modelled using random variables favouring the need of stochastic optimization algorithms such as Lyapunov optimization and stochastic approximation. Moreover, consumer interactions in a competitive environment implore the need of topics from game theory. The concept of Nash equilibrium is introduced and methods to identify such equilibrium points are laid down. Utilizing these notions, two research contributions are made. Firstly, a strategy for controlling heterogeneous energy storage units operating at different timescales is put forth. They strategy is consequently employed optimally for arbitrage in an electricity market consisting of day-ahead and real-time pricing. Secondly, energy storages owned by consumers connected to different nodes of a power distribution grid are coordinated in a competitive market. A generalized Nash equilibrium problem is formulated for their participation in arbitrage and energy balancing, which is then solved using a novel weighted Lyapunov approach. In both cases, we design real-time algorithms with provable suboptimality guarantees in terms of the original centralized and equilibrium problems. The algorithms are tested on realistic scenarios comprising of actual data from electricity markets corroborating the analytical findings. Real-time Integration of Energy Storage Sarthak Gupta General Audience Abstract Modern power system, which is responsible for generation and transport of electricity, is witnessing a lot of changes such as the increased adoption of wind and solar energy, promotion of electric vehicles, and ever increasing consumer demands. Amidst such developments, energy storage devices like batteries are being propagated as a necessary addition to the power system for its safe operation. This has been further supported by the decrease in prices of these devices over time. An effective assimilation of energy storage however, requires extensive mathematical studies on account of unpredictable renewable generation and consumer demands.This thesis focuses upon the preceding concern. To this note, two novel research contributions are made. In the first, an individual consumer is modeled who wishes to reduce his/her energy costs by simultaneously employing energy storages belonging to different technologies. In the latter, a more challenging multi-consumer interaction is reviewed where multiple end consumers wish to reduce costs while competing against each other over limited resources. In either of the cases, efficient algorithms are designed that are shown to produce desirable results over real-life data and have mathematically provable performance guarantees.